Chapter 3 Compaction and Diagenesis of Carbonate Sands

Chapter 3

COMPACTION AND DIAGENESIS OF CARBONATE SANDS A. H. COOGAN and R. W. MANUS

INTRODUCTION

Coarse-grained carbonate sediments compact under the pressure of overlying sediment, with increasing subsurface temperatures, through the mechanism of solution at the points of grain contact and diffusion of the solute into the pore fluid. Three main factors affect the compaction of carbonate sands: (1)the inherited factors such as packing, grain size and shape, and sorting; (2) the inhibitory factor6 which retard compaction, principally synsedimentary cementation; and ( 3 ) the dynamic factors which are connected with the subsurface environment, including the rate of loading, the geothermal gradient, the overburden pressure, the nature of the formation fluid, and others. The carbonate, mainly calcium carbonate, sands considered here are the grain-supported sands, the grain diameters of which are greater than silt size. Carbonate sands have been classified by a variety of schemes (Folk, 1959; Ham and Pray, 1962; Bissell and Chilingar, 1967; Bathurst, 1 9 q l b ) based on mineralogy, grain size (e.g., calcarenites), grain-to-matrix ratios (e.g., sparites), grain types (e.g., skeletal limestones) and framework or packing (e.g., grainstones). The terminology used for carbonate sediments and rocks in this chapter is not a complex one. The simplest terminology which conveys the relationships under discussion is used, so that the terms gravel, sand and silt are used as size terms and are used for unconsolidated and uncemented aggregates of those size classes regardless of mineralogical composition. Mineralogy is designated with an appropriate modifier quartzose, calcareous, aragonitic, etc. The limestones (the cemented and consolidated sediments) are named according to Dunham’s (1962) classification which stressed the grain-to-grain, grain-to-mud relationships more than other classifications. The limestones under discussion here are mainly the grainstones, that is the mud-free, grain-supported sediments and their lithified equivalents. In fact, most of those discussed are oolitic grainstones simply because more is known about them than about other compacted limestones of sand size. Until recently it was maintained that carbonate sands compact very little, if at all, owing to the widespread phenomenon of early cementation (Weller, 1959). This conclusion is still somewhat true because many

80

A. H. COOGAN AND R.W.MANUS

carbonate rocks are little compacted; however, some are very much compacted. The details of the compactive process and the exact parameters of the environment in which compaction proceeds are unknown. For the present, one must be satisfied with simple numerical approximations to the conditions of compaction; a comprehensive mathematical model is still lacking as are the data t o formulate one (see Chapter 3, Vol. 11, for discussion). Many carbonate sands begin as uncemented aggregates of organicallyderived particles with mineralogies unstable for most of the postsedimentary environments which they will occupy. Such sands can be altered and completely cemented in a geologic instant to stable assemblages of particles, with altered or preserved sedimentary fabrics, or can be preserved for tens of millions of years essentially unaltered. These quick-change artists, these houdinis of mineralogical composition contrast sharply with the conservative, little-changing clastic sands, mainly quartzose, discussed elsewhere in this book. The most fruitful approach t o understanding the compaction of carbonate sands so far has been (1) to compare and contrast them with clastic or artificial sands, (2) to compare the processes and environments of compaction, and (3)t o examine compacted rocks for evidence and characteristics of compaction. In this chapter the writers are concerned mainly with the grain-to-grain relationships. If carbonate rocks and sediments compact in the subsurface by a process of massive dissolution and transfer of the material away from the site, and the process leaves no trace of its existence or cannot be interpreted, then these types of rocks are disregarded here. Experimental studies of compaction have added certain limits and parameters which serve as tools for studying the process of compaction of carbonate sediments. It is important to know the natural grain fabric resulting from sedimentation, and before the compactive stress is applied during compaction, t o understand the changes in grain-to-grain relationships. This requires a consideration of the packing of particles, spherical and nonspherical, in regular and randomly ordered packs of single and multicomponent mixtures, because these relationships serve as the basis for quantitative estimates of the amount of cement derived from local-source solution and reprecipitation and the amount of compaction in selected carbonate grainstones. Once having established a practical starting point, i.e., a quantitative estimate of the grain fabric in the uncompacted state, approximations of the degree of compaction for limestones consisting of several kinds of particles can be made from measurements of grain volume and from grain-to-grain relationships.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

81

FACTORS AFFECTING THE COMPACTION OF CARBONATE SANDS

General statement

A survey of the literature and analysis of the evidence for compaction of carbonate sands and of the environment of compaction lead to the conclusion that there are three groups of factors which affect the compaction of carbonate sands. The first group of factors (Fig.3-1), termed inherited, reflects the original composition of the sand and includes particle size and shape, size and shape sorting, particle mineralogy, and sedimentary fabric (packing). The second group of factors, which inhibit compaction and thus can be termed inhibitory, are related to physically and biologically induced chemical changes during alteration. The alteration can be penecontemporaneous with sedimentation or subaerial exposure or can occur during other diagenetic stages. The main factors in this group are early cementation and dolomitization. The third group of factors, which are related mainly to the burial of the carbonate sand, are termed dynamic and include two main factors, Zhanges in overburden pressure and in temperature. Other dynamic factors are the rate of loading, length of burial time, composition and amounts of connate water and other fluids in the pore space, and pore fluid pressure. DYNAMIC

FACTORS

. Ovrrburdrn P r r a a u r r . Subaurfacr T r m p r r a t u r r

n .Durotlon of Burial S t r r a m . P o w Prraaurr .Porn F l u i d s

INHERITED

FACTOR

.Grain Sizr . Grain Shapr ' S i r r Sartlng . Shapr S o r t i n g Grain Packlng

.

INHIBITORY FACTOR P r r b u r l a l L I t h i f i c a t i on I. C r m r n t a t i o n 2 . D a l o m i ii r o t i o n

Fig.3-1. Factors affecting the compaction of carbonate sands.

82

A. H. COOGAN AND R. W. MANUS

In general, the inherited factors are mainly passive and affect the degree of compaction for a given compactive stress during burial. The inhibitory factors generally reduce grain-to-grain compaction, or if they are not present, their absence simply allows compaction to proceed more efficiently. The dynamic factors cause compaction of the sediment independently or in combination with other factors. Inherited factors The principal factors in the original carbonate sand which usually affect subsequent compaction are particle shape and size, size and shape sorting, fabric or packing, the homogeneity of packing, and particle mineralogy. Grain size The size of grains is important in compaction independent of its relationship to sorting, original porosity, or packing. In carbonate sands the grain size is clearly related to the grain type which, in turn, is related t o the biological or physical source of the particles. For example, coarse carbonate gravels cannot consist of planktonic foraminiferal tests which do not grow t o gravel-size proportions. As a result, the environments of origin of carbonate particles and of their deposition strongly control the sediment type and size. Grain size alone is important in compaction owing to the differences in surface area of particles. In addition, smaller particles give rise to larger surface areas per unit volume. As compaction proceeds, the thickness of the calcium carbonate dissolved from the particle at the point of contact, or by a through-flowing acidic fluid or fluid undersaturated in calcium carbonate, is relatively greater for small particles than for larger ones. If in the course of compaction, a 0.005-mm thick slice of aragonite is removed from a grain, this amounts t o a 2% change in radius for a medium size grain (diameter = 0.5 mm), whereas it amounts to a 10% change in the radius of a very fine sand (diameter = 0.1 mm). As a result, given other constant conditions, it would be expected that finer sized sands should compact more than coarser ones. There is no statistical evidence known to the authors that this is true for carbonate sands, probably because so many other factors are involved that the effect of grain size is combined with them in a still undecipherable mix. Grain size also plays a role in cementation which in turn affects compaction. Grain shape Particle shape is important in compaction and is critical in any consideration of the original packing configuration of the sediment as a result of sedimentation. Almost all analyses of the packing of materials have

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

83

been done with spherical particles (Graton and Fraser, 1935; Brown and Hawksley, 1945; Bernal and Finney, 1967; Ridgway and Tarbuck, 1967; Coogan, 1970; Morrow, 1971; Rittenhouse, 1971; and others). These experiments and analyses are generally based on the ideal geometric arrangements of spheres which correspond to one of the regular packing configurations or to some unordered natural arrangement (Fig.3-2). It is the natural precompaction packing configuration of the sediment which must be quantified in order t o evaluate the extent of subsequent compaction (Coogan, 1970). This information has been lacking for most carbonate sands. With the exception of a small contribution by Graton and Fraser (1935) on the packing and porosity of mica and minor consideration by Dunham (1962) on the packing of leaf-like and other irregular algal particles, almost no information has been available on the packing of irregular-shaped particles either in single- or varied-shaped particulate sands. This applies to Recent and ancient carbonate rocks, with the exception, of course, of nearly spherical oolitic sands. Importance of the problem of varied shapes in carbonate sands was recognized by Maiklem (1968) who studied the hydraulic properties of bioclastic carbonate grains. He recognized four basic shapes: blocks, rods, plates, and spheres. Dendroid and irregular particles can be added to the list. While he does not consider the effect of these shapes on subsequent compaction of the sediment, his paper is useful in recognizing the complexity of the particle shapes in carbonate sands. The shape of carbonate particles is a reflection of the (1) abundance, availability and type of contributing organisms; (2) growth-size distributions of the organisms; and (3) broken size distributions of the contributing skeletal materials. The latter depend on the mechanisms of size and shape reduction (crushing, abrasion, chipping, solution or decay-disintegration) and on the skeletal structure of the material (internal cavities, anisotropism of hardness, and crystal bundles). These inherited parameters of the shape distribution change with the biological province, the local environmental setting of the carbonate environment, and geologic age. To this list of parameters can be added the later history of hydraulic modification of the grain

Fig.3-2. Regular and random packing of uniform spheres, displayed as a monolayer. A. Regular open cubic packing. B. Regular close hexagonal (rhombohedral) packing. C. Unordered packing.

84

A. H. COOGAN AND R. W. MANUS

shape distribution owing to transport of the particles as sediment to its site of deposition, method of transport (bottom traction, saltation, suspension), length of time, distance and selectivity of transport of the original shape distribution due to the competency and capacity of the agent, and the rate of supply of the material. The subsequent mode of deposition, whether by dumping or gradual (i.e., rate of sediment accumulation), partial loss of the load, and post-depositional reworking by benthic animals and plants are additional processes which may affect the final shape distribution of a carbonate sediment. In general, carbonate sands are mixtures of two or more of the shapes discussed by Maiklem (1968). There are exceptions, however, such as nearly single-component algal sands and gravels and crinoidal ossicle sands, both of which may form from nearly single-constituent, preservable biotic populations that are barely transported. It was shown that the shape directly affects the settling velocity. When compared to calculated velocities based on mineral density, this velocity may be as much as four times slower. Block-shaped grains such as coral and coralline algae show the least differences between calculated and actual settling rates. Plate-shaped grains show the greatest differences and these differences in settling velocity, which certainly must be reflected in the sedimentary structures of the undisturbed carbonate sand, directly contribute to variations in the sedimentary packing arrangement (Fig.3-3).

Fig. 3-3. Diagrammatic illustrations of random packing of varied carbonate grain shapes, randomly cut. A. Spheres. B. Rods (whole gastropod shells). C. Plates (pelecypod valves). D. Irregular plates (plate algae). Not to scale.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

85

In comparison to quartzose sand particles, carbonate particles are extremely diverse in shape and the parameters of shape, i.e., sphericity and angularity. Many grains in quartzose sands have sphericities of 0.7-0.8 (Rittenhouse, 1971) or can be described as spheroidal in contrast to the disks, blocks, plates and rods of common carbonate grains. Carbonate particles tend t o approach sphericity less rapidly than quartzose grains in their respective environments, but tend to round faster owing to their lower hardness. Although carbonate grains before abrasion are generally smooth, occasionally they have angular ribs or spikes. They are rapidly rounded in environments with any appreciable wave motion or abrasion. In general, the diverse shape of carbonate particles is more important in packing, and hence in compaction, than the angularity of the grains. Finally, the biological processes that produce carbonate particles with abundant internal cavities (e.g., gastropods, foraminifers, cephalopods) produce a sediment which differs considerably from that made of clastic particles or carbonate particles with a more solid construction (e.g., corals, crinoid ossicles). The former sediment may have a very high internal (intraparticulate) porosity which, when filled with sediment or fluid, could modify the effect of overburden pressure, at least at shallow and moderate depths.

Size sorting

Sorting in sedimentary rocks is an important character of the fabric of the sediment that modifies the effects of compactive forces. In addition to variations in grain size usually described by the terms weEl to pooriy sorted, the distribution of sorting patterns in the sediment, or the arrangement of clusters having varied sorting, is important in understanding the effect of compaction on a grain sediment. This is especially true for the phenomenon known as bridging, which occurs where large grains receive the bulk of the vertically applied stress and small grains between them are less affected. Other kinds of sorting, i.e., sorting by shape, sorting of clusters of grain sizes, sorting of fabrics (Morrow, 1971), and sorting of intraparticulate pore space, are also important factors affecting compaction. Familiarity with the concepts of regular and irregular packing in sediments, however, is required before discussing them.

Packing The term packing refers to the spacing or density pattern of grains or pores in a sediment or rock. It is in part synonymous with the term fabric. Fabric, however, also refers to larger-scale sedimentary features such as crossbeds and burrowed structures. The difference can be illustrated with an example: A cross-bedded sediment composed of grains with homogeneous,

86

A. H. COOGAN AND R. W. MANUS

irregular close packing may have an average grain volume of 63%and pore volume of 37% calculated over the observed surface of the sample. An intensely burrowed sand may have the same packing and average grain volume, but obviously different gross fabric. Original packing is important in compaction because it represents the natural arrangement of the grains and pores on which the compactive stress will be applied. Regularly ordered, grain-supported sands of spherical particles may have grain volumes as high as 74.1%for regular rhombohedral packs and as low as 52.4%for regular cubic packs. Naturally sedimented sands appear to have unordered average grain volumes between 50% and 70%. Grain-supported carbonate sands of predominantly nonspherical particles with poor size sorting, such as branching algae, may have grain volumes considerably below the loosest spherical cubic packing of 52.4%, approaching 40, 30, or even 20%.I t should be emphasized that the lower the initial grain volume, the more the sediment must be compacted; therefore, chemical compaction should proceed more readily in sediments with initial close-packing configurations. A detailed discussion of packing configurations, measurements of packing, and variations in packing is presented in the following sections. Mineralogy Chemical and mineralogical studies of Modern and Pleistocene shallowand deepwater carbonate sediments (Chave, 1962; Chave et al., 1962; Friedman, 1964, 1965; Milliman, 1966; Matthews, 1967, 1968; Land and Goreau, 1970; Benson and Matthews, 1971; Huang and Pierce, 1971) show that these sediments are composed predominantly of metastable carbonate minerals (high-magnesium calcite and aragonite) with lesser amounts of the more stable low-magnesium calcite. The original mineralogy is largely a direct product of the metabolic processes of the plants and animals which contribute the skeletal grains to the sediment or, and to a much smaller extent, of physicochemical processes. Most of the magnesium present substitutes for calcium by solid solution in the calcite lattice in such common marine animals as foraminifers, sponges, coelenterates, holothuridians, crinoids, brachiopods, molluscs, and ostracodes. In contrast, the skeletal parts of madreporarian corals, many pelecypods, gastropods and cephalopods are aragonitic, as is the principal physicochemical grain carbonate sediment, i.e., oolitic sand. Thus carbonate sand differs greatly from clastic sand in being more varied in particle size and particle shape. It is Jess varied in particle mineralogy, but has high differential particle solubility owing t o different proportions of skeletal contributions t o the sediment. Major element composition varies with carbonate grain size. Analyses by Taft and Harbaugh (1964) of southern Florida sediment show that there

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

87

is a progressive decrease of the calcium/magnesium ratio with decreasing grain size. Samples from the Bahamas (Taft, 1967) show that sand-size carbonate particles tend to have the highest proportion of aragonite. In contrast, the Recent lime muds of southern British Honduras (Matthews, 1966) contain as much as 44% high-magnesium calcite derived from skeletal disintegration. In summary, the mineralogy of carbonate particles is controlled by biogenic productivity and subsequent alteration. The striking mineralogical difference between modern sediments and ancient carbonate rocks is that the latter commonly have little of their original mineralogy. Thus the lithification* process usually involves changes in particle mineralogy. Minor element composition, however, is probably not a major factor in lithification during compaction.

Inhibitory factors Certain factors can inhibit the compaction of carbonate sands so that at moderate depths of burial, e.g., between 1,500 and 5,000 m, there may be little physical evidence of compaction even though the rock is lithified. The cause for the lack of compaction in carbonate sands is generally considered to be the result of early cementation (Weller, 1959), which may preserve loose grain packing and uncrushed fragde shells even at great depths. Early lithification* may occur: (1)penecontemporaneously with sedimentation, even as part of the sedimentary process; (2) during postdepositional periods of subaerial exposure of the carbonate terrain owing t o slight regional uplift or low stands of sea level; ( 3 ) to various submarine and shallow subsurface changes in pore water composition related to a fresh-water lens or artesian upwelling; or (4) to biological activity. The extensive phenomenon of early lithification of carbonate sediments (Bathurst, 1971b) led many (Weller, 1959) to conclude that most limestones show no evidence of compaction. Lithification before burial tends to add strength to the sedimentary unit, fills the pores with cement, and removes fluids which might aid in dissolution. The principal processes of early lithification include cementation and dolomitization.

Cementation An extensive literature on diagenetic cementation of’ carbonate sands (LeBlanc and Breeding, 1957; Bathurst, 1958, 1959, 1971a, b; Friedman, 1964; Pray and Murray, 1965; Larsen and Chilingar, 1967; Matthews, 1967; Bricker, 1971; and many others) reveals the source and characteristics of the *Lithification: that complex of processes that converts a newly deposited sediment into an indurated rock.

88

A. H. COOGAN AND R. W. MANUS

cement, the driving force and conditions of cementation, and the chemical reactions which illustrate that many sedimentary and diagenetic environments favour dissolution, reprecipitation, and cementation. Purdy 's (1968) summary emphasizes the diagenetic environment. He showed that carbonate sediments are transformed into limestone in subaerial, submarine, and subsurface environments. While not extensively distributed or massively pervasive, cementation, principally as minor pore filling, occurs in submarine shallow tropical waters, for example, in the oolite and grapestone bars of the Bahamas (Purdy, 1968; Taft, 1968), and in the deep sea as reported by Milliman (1966), Fischer and Garrison (1967) and Thompson et al. (1968). In contrast, subaerial processes are of considerable importance. They involve solution, replacement and cementation, which affect major portions of the carbonate terrain of a sedimentary basin, principally through the efficacy of percolating meteoric water in the vadose zone where solution and replacement occur together (Matthews, 1967; Benson and Matthews, 1971; Bricker, 1971). It is difficult to explain the source of large amounts of cement which are present in carbonate rocks. As Bathurst (1971b) has pointed out, petrographic evidence commonly suggests that cementation of limestones begins early, i.e., before compaction, and that about half of the volume of calcium carbonate in the rock has come from a source outside that of the immediate sediment. Redistribution of calcium carbonate by dissolution and reprecipitation is inadequate to fill the pores of most carbonate sediments in the uncompacted state. The possible sources of cement are the local dissolution of aragonite, influx of cement in pore-filling sea water, and pressure sohtion in subsurface environments. Cementation in the subsurface below the depth of a few meters is poorly understood. The factors which affect the chemical equilibrium are many, although temperature and pressure appear to dominate. From the geological standpoint, it appears that a given carbonate sediment may reach a state of complete cementation at any depth of burial between the surface and that at which metamorphism begins. Once the sediment is completely cemented, compaction through grain-to-grain adjustments is unlikely.

Dolomitization

Recent dolomitization at or near the surface of sediments has been described by many investigators since the early 1960's from South Australia, the Persian Gulf, Florida, and the Bahamas (Alderman and Skinner, 1957; Shinn, 1964; Deffeyes et al., 1965; Shinn et al., 1965; Atwood and Bubb, 1970; Butler, 1971). Penecontemporaneous dolomitization is part of the process of sedimentation on carbonate tidal flats. In the most general way, this type of dolomitization occurs as the result of influx of magnesium-

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

89

bearing tidal waters into the carbonate sediment, concentration of the pore water by evaporation, and precipitation of calcium carbonate or calcium sulfate. The process results in a magnesium-rich brine with the consequent substitution of the magnesium for calcium t o form dolomite in the sediment over thousands of years. The process may produce near stoichiometric dolomite or may partly dolomitize limestones of any texture t o depths of a meter or so and over areas of several thousands of square kilometers. The second major process of extensive dolomitization is held to be a result of movements of magnesium-rich brines by gravity flow from an area of brine concentration, such as a tidal flat or restricted pond, into contemporaneous or older carbonate rocks of any texture. This process of seepage dolomitization has been proposed and partially documented in West Texas, the Persian Gulf, Bonaire in the Antilles, and Great Inagua Island, Bahamas (Adams and Rhodes, 1960; Chilingar and Bissell, 1961; Deffeyes et al., 1965; Bubb and Atwood, 1968; Murray, 1969; Butler, 1971). Both of these processes of dolomitization can produce a lithified dolomite or dolomitic limestone extensively cemented by calcite, with a substantial Stratigraphic thickness and areal extent. The rigid, cement-bound framework of the resultant rock should resist compactive stress equivalent to many hundreds of meters of overburden.

.

Dynamic factors The dynamic factors, those causing or contributing to compaction of the carbonate sand, are the most difficult to assess independently in the field, because there are almost no data on natural compactive processes below depths of about a meter. Experimental data on compaction of coarsegrained sediments is available for some carbonate sands (Fruth et al., 1966; Ayer, 1971) and are 'discussed later. In this section, the various factors directly affecting compaction are outlined together with their independent roles. The two principal factors are: (1)the increase in overburden pressure with depth (also referred t o as increase in depth of burial, maximum effective stress or applied load, depending on the locale and the author), and (2) the increase in temperature with burial, generally discussed in terms of geothermal gradients or in the laboratory as measured temperatures. Other factors include the rate of loading, the length of geological time or measured time, the composition and movement of interstitial fluids, and the static or dynamic pore pressure. Maximum effectiuestress or overburden pressure The maximum effective stress to which a stratigraphic unit or individual grains are subjected in a sediment or rock with pore space can be expressed

90

A. H. COOGAN AND R. W. MANUS

according to the formula presented by Terzaghi (1936) as: a = S - p , where S is the total load, p is the pore-fluid pressure and a is the effective or matrix stress. The formula draws immediate attention to the close relationship between the effective load (grain-to-grain stress of the sediment) and the buoyancy provided by the pore fluids. For compaction t o proceed, the pore fluid must migrate as a gas or fluid out of the sediment in response to the applied pressure. The fluid usually can migrate until the sediment is matrixor grain-supported. Most carbonate sands reach this stage at the end of sedimentation stage. Further fluid migration is necessary as the grains are crushed or welded together and as cementation occurs in the diminished pore space. Very little data are available on the pore pressure in rock units or in sediments undergoing compaction, after the grain-supported state is reached. Generalized numerical estimates of the total load (Weller, 1959) show that there is increasingly greater total load with depth. Below a depth of about 175 m and porosity of less than 3776, the weight of pore water must be considered. Above a depth of about 175 m, the effect of hydraulic lift is assumed to be completely effective. (See Chapter 2 for detailed discussion.) In carbonate rocks the situation is more complex than in sandstones owing t o early cementation. There are, however, a few instances, for example, in the Bikini and Eniwetok drill holes (Schlanger, 1963), where some intervals of carbonate grain sediments, which are buried as deeply as 900 m, exhibit little mineralogical alteration to low-magnesium calcite and little cementation, such that the sediment recovered is friable and not much compacted. The JOIDES program of drilling on the Blake Plateau off northeastern Florida included recovery of cores of unconsolidated, essentially uncemented and uncompacted, planktonic carbonate sands and muds as old as Eocene and Paleocene (Schlee and Gerard, 1965; Bunce et al., 1966). In the more common case, for example, in the Bahamas and Florida where carbonate rocks of Pleistocene and Late Tertiary age crop out, the sedimentary column is cemented from the surface downward, owing to early diagenetic cementation in the submarine, vadose and phreatic zones. In such cases, consideration of the effects of present-day overburden load on grain relationships is made substantially more difficult. Thus a further study of the effect of compaction progressively deeper in the stratigraphic column by drilling and use of better tools for evaluating compaction is needed (see Chapter 7). Extrapolations of data on the pressure due to overburden load have been made (Weller, 1959; Maxwell, 1960) for depths far in excess of those penetrated by current drilling methods. Maxwell (1960) estimated that pressures at a depth of 14,000 m should approximate 3,100 kg/cm2, an

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

91

extrapolation slightly beyond Weller's greatest depth. Although the reported data and extrapolations are for sand-shale sequences, the overburden load for wholly carbonate sections should be of similar magnitude. Temperature changes From the surface downward to a depth where all porosity has been eliminated by compaction, the temperature may be expected to increase by several hundred .degrees centigrade (Weller, 1959). Experimental work by Sippel and Glover (1964) has shown that the effect of temperature on the solubility of calcium carbonate is much greater than the effect of pressure at moderate depths. For example, at a depth of 1,000 m and a concentration of 220 mg/l calcium carbonate, a 10% increase in temperature at constant pressure would produce a decrease of about 30 mg/l in the concentration. As depth increases, the effect of change in pressure on calcium carbonate solubility becomes more pronounced. It is not clear how these changes affect the micro-environment of grain-to-grain, grain-to-pore and grain-to-cement relationships in a buried sediment, because the temperature probably affects the whole of the sediment equally, whereas the pressure is concentrated at the points of contact. In an extrapolation of Sippel and Glover's (1964) and Sharp and Kennedy's (1965) data, Bathurst (1971b) projects the solubility of calcite with increasing temperature between zero and 600" C for pressures of 412 kg/cm2 (400 bars) t o 1,712 kg/cm2 (1,400 bars). The graph (Bathurst, 1971b) shows that a calcium carbonate solubility of 11mg/kg of solution would be reached at about 260°C for a pressure of 200 kg/cm2, whereas a temperature of over 400°C would be required to reduce the solubility to the same level at a pressure of 1,712 kg/cm2. At a depth of 6,000 m, the solubility should be reduced a whole order of magnitude, hence reducing the rate of dissolution and retarding compaction. On the other hand, even at depths of 8,000 m and temperatures of about 4OO0C, calcium carbonate is still soluble and in the presence of favorable conditions (fluid movement, unstable mineral species) it would be expected that compaction probably would continue. In general, one might expect that in areas of low geothermal gradient compaction should proceed more rapidly. The most readily obtainable data on geothermal gradients and their effect on compaction are provided by Maxwell (1964J for clastic sedimentary basins. His data show that temperature is an important variable, but one about which there is little reliable information, especially at great depths. Bottom-hole temperature readings probably give reliably consistent results, at least on the average, to depths of 5,000-7,000 m. For greater depths the published temperature data are scarce. It is useful to consider the range of geothermal gradients encountered by Maxwell (1964) in his study of quartzose sandstone porosity at depth. The range is between 7" and

92

A. H. COOGAN AND R. W. MANUS

12"C/350 m. He found that clastic sands had lower porosities in areas of high geothermal gradients, probably reflecting the change in solubility of quartz with temperature and increased cementation.

Rate of loading The rate of application of stress to the compacting sediment is one of the two time factors important in Compaction. As discussed previously, Terzaghi's (1936) formula (a = S - p ) shows that compaction cannot proceed much unless the pore fluid is able t o migrate. Only when the rate of loading is slow in terms of fluid migration will there be a compaction equilibrium at all times. The rate of fluid movement is directly related to permeability. In the case of carbonate sediments and rocks with inter- and intraparticulate pore space, the degree of communication between the pores is important. Low permeability may result in long periods of time for attainment of equilibrium process, because of its influence on the expulsion of large quantities of water where adjacent pressure differentials are small and friction is relatively great. In the case of rapid loading, longer time is needed for the compactive system t o equilibrate. In carbonate stratigraphic sections, the likelihood of occurrence of extremely rapid loading is not great. The average rate of deposition in carbonate sedimentary basins may be as rapid as in clastic basins; for example, more than 5,000 m of shallow-water sedimentary rocks in the Bahamas have accumulated since the Middle Cretaceous at a rate of deposition at least equal to the rate of subsidence. It probably is true, however, that locally, for example off the present Mississippi River mouth, sediment loading in this clastic depocenter is much more rapid than in any carbonate sedimentary province. Scarcity of overpressured, shallow-depth sedimentary sequences in carbonate areas tends t o suggest that loading was either sufficiently slow or that communication and egress of fluids was great enough, so that equilibrium was maintained in Carbonate sand bodies under compaction. Length of time of burial The duration of burial is the second time factor affecting compaction. Like the rate of loading, duration of burial is related t o the ability of the sediment or rock to expel pore fluids and provide space for the collapse of the grains or movement of grains closer together and for cementation. It is important to note that duration of burial is not the same as the geological age of the carbonate sand. Instead, it is the duration of burial as the sediment is subjected t o a given stress. For ancient rocks this is difficult to evaluate even if the history of the sedimentary basin is known in some detail. Experimental work in one petroleum company laboratory has suggested that

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

93

for quartzose sandstones with the same packing, pore fluid pressure, geothermal gradient, salinity, pH of formation water, and average grain size, an increase in duration of burial, for example from 5 to 500 million years, should reduce the porosity from 35%to 5%,with a corresponding increase in the amount of compaction.

Formation fluids The existence of varied interstitial fluids in the pores of incompletely compacted sediments, especially fluids which wet the surfaces of the grains, probably plays a role in the process of compaction. There exists a potential for retardation of compaction through pressure solution in carbonate sands which have hydrocarbon-wet surfaces, because the relative permeability to water would be small and the consequent movement and uniform distribution of cement would be difficult. Oil-wet particle surfaces could substantially reduce the potential for solution of cGbonate minerals and the redeposition of calcium carbonate as cement in the immediately adjacent pore space. The actual effect of hydrocarbons on cementation in the pore space of carbonate rocks is unknown. Oil and gas have been suspected of controlling or maintaining open pore space in clastic sandstones in many fields and in the Smackover Jurassic oolite, but the documentation usually has been questionable. Close association among presence of oil, higher than expected porosity, and slight evidence of compaction does not necessarily mean that the hydrocarbons have preserved the porosity; in fact, it may mean only that the high porosity has served to localize the oil. It has been suggested that the presence of gas does not help preserve porosity in the clastic sediments in south Louisiana (Atwater and Miller, 1965). There, an average decrease in porosity of 1.2% per total volume for each 350 m of burial, observed in 3,000 gas reservoirs, indicates little effect of gas presence on porosity. This rate of porosity decrease was essentially that observed for the water-filled sands, too. From the general knowledge on pressure solution and reprecipitation now available, it does seem unlikely that oil or gas would greatly inhibit pressure solution as long as a thin film of water surrounds each grain, which is the case in the typical water-wet oil sand, whether quartzose or carbonate. In a study of the differences in compaction of sands in petroliferous and water-bearing strata of Early Cretaceous reservoirs of the Ust' Balyk and West Surgut Fields, Western Siberia, Zaripov and Prozorovich (1967)" compared the total number of impressed contacts between grains of quartz, feldspar, and rock fragments. Their findings show markedly higher numbers of impressed contacts for quartz grains in the water-wet beds over those in *For other references see Chapter 3, Vol. 11, o n sandstone diagenesis.

94

A. H. COOGAN AND R. W. MANUS

the oil-wet beds at the same depths (2,100-2,300 m). They found the differences t o be statistically significant. Although they concluded that in the oil-saturated sandstone the dissolution of quartz ceased or slowed down after the reservoir had been filled, compaction of the reservoir sand continued chiefly through plastic deformation of rock fragments so that both oil- and water-wet beds continued to compact. In addition, they noted that coarser sands showed more effects of pressure solution owing to greater pressure per grain per contact and the greater permeability of the sediment. Formation waters may increase in salinity dramatically with increase in depth of burial. In general, the diagenesis of connate sea water may involve a reduction in the pH locally owing to the decomposition of organic matter not previously oxidized near the surface, resulting in a slight increase in magnesium and calcium ion concentration of the solution. With further burial and compaction, some of the salty water should be expelled. Where porous and permeable carbonate units are interbedded with other sediments, this could mean expulsion of brines into or through the carbonate sediment. Review of the literature indicates there are no data on the relative amounts of compaction of carbonate sands interbedded with clastic shales and those which are part of massive, thick carbonate units. Influx of substantial amounts of low-pH water, as the result of (1) exposure to a fresh-water supply at the surface, (2) upwelling artesian flow into a carbonate sand body, or (3) expulsion from a clastic sediment into a carbonate body, would greatly affect the mineralogy of the original carbonate sediment tending to convert the carbonate particles to lowmagnesium calcite. Maintenance of a static, low-pH fluid in the pores of a carbonate sediment is unlikely. Lack of experimental work on compaction of carbonate sands in the presence of moving acidic waters hinders further analysis, but one might speculate that a through-flowing acidic fluid would tremendously enhance compaction by accelerating massive dissolution. Summary o f geological factors favoring compaction o f carbonate sands Considering the dynamic, inhibitory and inherited factors affecting compaction, it appears that the compaction of a carbonate sand is more likely in geological settings where the following conditions prevail, recog nizing the fact that if compactive stresses are high enough, deformation will occur regardless of other factors: Dynamic: (1) Burial is deep and effective stress is high. (2) Geothermal gradient is low. (3) Rate of loading is low.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

95

(4) Pore pressure is low in a static fluid, or there is sufficient permeability to allow fluid expulsion. (5) Duration of burial under stress is long. (6) Through-flowing fluids are acidic and undersaturated with respect t o calcium carbonate. (7) Grain surfaces are water-wet.

Inhibitory : (1) Early cementation is minimal. (2) Dolomitization is minimal. (3) Diagenesis and alteration to low-magnesium calcite is minimal. Inherited: (1) Average grain size is small. (2) Grains are moderately to well sorted: (3) Packing, as a result of deposition and penecontemporaneous processes such as burrowing, is the most close packing possible for the particular sediment. (4) The mineralogy of the carbonate sediment is of the least stable kind for the subsurface environment. PRECOMPACTION GEOMETRY - PACKING

General pertinency of packing to compaction The critical problem facing the geologist who desires to measure the amount of compaction, which has taken place in a grain carbonate rock, is to determine the point of zero compaction. He must choose the numerical value which represents the starting point or the packing arrangement of the loose, uncemented sediment before any adjustment, rotation, slippage, peeling, fracture, or pressure solution has occurred. Two principal sedimentary characteristics which affect the packing are the shape of grains and their geometric arrangement. Almost all theoretical and experimental studies of packing have used spherical particles or points in space (Kelvin, 1887; Graton and Fraser, 1935; Marvin, 1939; Matzke, 1939; Brown and Hawksley, 1945; Bernal and Finney, 1967; Ridgway ana Tarbuck, 1967; Coogan, 1970; Morrow, 1971; Rittenhouse, 1971). The principles are well understood and, for geologists interested in sediments, have been presented at length by Graton and Fraser (1935), who dealt with the regular 'packing of spheres and progressed to more irregular, random packing arrangements. Naturally, there is criticism of using spheres as model

96

A. H. COOGAN AND R. W. MANUS

grains (see Chapter 2). Morrow (1971) has extended the analysis to systems of heterogeneous and homogeneous regular and irregular packing. Rittenhouse (1971) has related the effects of compaction on certain packing arrangements to the maximum degree of cementation to be expected from local-source cement during the compactive process. In this section, the writers present an outline of the regular packing configurations and then the unordered ones, which can be derived from them, as a basis for choosing methods of measuring packing and hence compaction in carbonate sands. Regular packing of spheres By far the most complete analysis of regular packing of spheres is that of Graton and Fraser (1935), who summarize the configurations which are stable against the force of gravity acting alone. There are two types of layers, the square and simple rhombic (Fig.3-4A, B), and there are three simple ways of stacking these layers one on top of another (Fig.3-5A-F). Because two of the three ways of stacking square layers are identical to, but differently oriented from, two of the three ways of stacking simple rhombic layers, there are six fundamental regular arrangements. Two of these six arrangements repeat as to form (grain volume, porosity), but differ in symmetry and hence tortuosity and permeability. The six regular geometric arrangements of four spheres are named according t o crystallographic terminology the cubic, orthorhombic and rhombohedral packing of square layers (Fig.3-5A-C) and the orthorhombic,

Fig.3-4. Plan view of two types of regular packing layers. A. Simple square layer. B. Simple rhombic layer.

COMPACTION AND DIAGENESIS O F CARBONATE SANDS

97

A

B

D

E

Fig.3-5. Six regular packing configurations of uniform spheres; view from above. Cases 1 to 3 are arrangements of square layers (Fig.34A). Cases 4 to 6 represent similar offsets for stacks of rhombic spheres (Fig.3-4B). A. Cubic, case 1; four spheres sit directly above four other spheres. B. Orthorhombic, case 2; four spheres offset one-half sphere in one direction with regard to underlying spheres. C. Rhombohedral, case 3; four spheres offset one-half sphere distance in two directions with regard to underlying spheres. D. Orthorhombic, case 4. E. Tetragonal, case 5. F. Rhombohedral, case 6.

tetragonal and rhombohedra1 packing of simple rhombic layers (Fig.3-5WF). For each of these six cases of regular packing, there is a fixed ratio of grain to pore space and each has a definite stability related to the number of grain contacts below a given grain (Table 3-1). In addition, the six packing types can be characterized in terms of coordination numbers and the number of other grains touched by an arbitrarily chosen central grain. For example, in' case 1 (cubic packing), where uniform spheres sit precariously atop other spheres (Fig.3-5A), each central sphere touches six others

98

A. H. COOGAN AND R. W. MANUS

TABLE 3-1 Grain and pore volume, stability, and number of grain contacts of the six regular packing types

Packing type

Grain volume

(a)

Pore volume

f%)

Stability

Number of contacts by each grain below

Case 1 , cubic Case 2, orthorhombic Case 3 , rhombohedral Case 4 , orthorhombic Case 5, tetragonal Case 6, rhombohedral

52.36 60.46 74.05 60.46 69.81 74.05

47.64 39.54 25.95 39.54 30.19 25.95

low medium high low medium high

1

all sides 6 8

2 4

12

2

10

1

4

8

12

(Fig.3-6A). When displayed as a monolayer, four of the six are in the same plane and two in another plane. For case 6 (rhombohedral packing), where spheres lie in the “holes” between underlying spheres, each central sphere touches eight others (Fig.3-6B), but only six are shown when displayed as a monolayer. In the two respective cases, the grains may be said to have six- or eight-fold coordination. Coordination numbers, which are common in crystallographic terminology, are almost impossible to derive for naturally compacted sediments and rocks. Their packing must be described in other less precise terms, partly because the packing of natural grains is seldom regular to any appreciable extent and partly because of cementation. The porosity of regular packing configurations is fixed and directly related to the geometric arrangement of the pack. Because the sorting of grains of the six regular cases is regular, the sorting of their pores is also regular. In cubic packs the pores appear as diamond shaped in the display of spheres as a monolayer. Graton and Fraser (1935) made extensive determinations of the geometry of the pore space, including serial sections of the pores in the regular packs (Fig.3-7). By way of contrast, two different monolayers are shown (Fig.3-7), which emphasize pores and grains at different positions for the case 1 and case 6 packing. It is clear that the space in rhombohedral packs is not only less (25.9% porosity versus 47.6% for cubic packs), but the geometry of the pore space is markedly different. In rhombohedral packs, there is less grain surface area exposed t o wetting fluids and the capillary pressure and surface tension for the whole pack are higher. All of these characteristics affect the subsequent compactive forces as they relate t o solution of the grains under pressure and the maintenance of pore pressure:

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

99

Fig.3-6. Sphere contacts in regularly-ordered packing of uniform spheres, shown as a monolayer. A. Cubic packing; each sphere touches 6 others, 2 are not in the plane of the monolayer. B. Rhombohedral packing; each sphere contacts 8 others, 2 are not in the plane of the monolayer.

Fig.3-7. Geometry of pore space in cubic and rhombohedra1 packing seen as though the solid spheres were removed. A Cubic pores. B. Rhombohedral (case 6, Fig.3-5F) pores. Left-hand figure of each set represents the three-dimensional pore space in the unit cell. The right-hand figures represent different cross sections of this three-dimensional pore.

100

A. H. COOGAN AND R. W. MANUS

Stability of the pack is directly related to the number of grain contacts vertically and laterally made by each grain in the pack. Any sphere acted upon by gravity and supported from below requires at least three points of support in order t o attain equilibrium and a fixed and stable position. Of the six packing types, only the two rhom-bohedral ones (cases 3 and 6, Fig.3-5C and F) fulfill this requirement because only in these two cases do spheres of any layer have three or more points of support from those of the underlying layer. In the orthorhombic and tetragonal (cases 2 and 5, Fig.3-5B and F) packs, each sphere receives support from below at only two points and is balanced in the cusp between the two underlying spheres except for the lateral support afforded by neighbors in its own layer. In the least stable configuration, the cubic and orthorhombic packs (case 1 and 4, Fig.3-5A and D), each sphere is perched tenuously on the pinnacle of a single underlying sphere and requires lateral support from neighboring spheres. In general, of any two uncemented packs, the one that is more stable has the lower porosity and the higher grain volume. In the more stable pack each sphere touches the largest number of adjacent spheres and the vertical spacing between layers is smaller. Such stable packs should undergo the least amount of consolidation by their own weight or any minor disturbance such as slight sedimentary loading near the depositional ,surface. In addition, stable packs should have to adjust the least amount in attaining complete compaction- that is, the configuration with 100% grains and no pore space - through the application of stress.

Random packing of spheres Definition and origin There are only two explicitly definable states o f a packed bed. The first is when the bed is completely ordered and the grain volume is at a maximum for the configuration. The second is when the grain volume is at a minimum so that any decrease in grain volume would result in a cloud of separate particles, that is, a quick sand. Intermediate states are forms of random packing. The random packing of spheres or other grains occurs under natural conditions of sedimentation. No examples of the six ordered packing types have been reported for naturally sedimented carbonate grain sediments to the knowledge of the authors, even for nearly spherical oolitic sediment, with the exceptions of possibly small clusters of ordered grains of cubic or rhombohedral packing occupying small regions of an otherwise randomly-packed sand. These random packs may exhibit either less or more complicated kinds of orderliness than characterize the six simple cases or may contain combinations of ordered and unordered packing. They cannot, however, be characterized as lacking in order entirely, although disorderly

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

101

packing appears generally in random packing. The stability of random packs is intermediate, and their variable grain volume and porosity are commonly described in terms of averages for the whole pack of sample, mainly for convenience and because averaging measurement techniques are used. It has been shown (Graton and Fraser, 1935) that random packing will occur in simple cases where spheres are poured freely and fortuitously into a container. Vigorous jarring might cause slight reordering, especially where the initial packing is loose; however, random packing, which is fairly typical of natural sands, is usually stable. On the average, random packing is statistically reproducible. It is important to note that random packing, once established in the midst of a large sand body, cannot be completely eliminated or translated into one of the more systematic types of packing by any amount of mere jamng. This is because the spheres are locked into position and openings, which are more than large enough t o accommodate a whole sphere, may be left unfilled by the manner of sphere accumulation. These openings are protected by arching of the surrounding spheres and cannot be filled without the lifting of some stable individual grains. Hence, random packing, which is not the most stable theoretical configuration, has a great natural stability such that carbonate sands can become rapidly and permanently fixed in a randomly-packed state by early cementation around grain boundaries near the depositional surface. Chance packing, the natural fortuitous combination of systematic regions of ordered packing, mainly of the case 6 and case 1 (Fig.3-5F and A) configurations commonly occurs with intervening random, truly orderless zones in carbonate sands of nearly uniform grain shapes and sizes, such as oolitic sands. No direct quantitative data are available on- the relative amounts of such chance packs, but indirect evidence supplied by Graton and Fraser (1935) and by Morrow (1971) indicates that chance packs do occur in natural sands. Density of random packs The grain volume of varied types of random packs is important to determine as a basis for establishing the zero point of compaction. Samples of spherical grains prepared in the laboratory in relatively large containers with more than ten million grains, or collected from various sedimentary environments, have measured porosities of about 37% which corresponds to grain volume of about 63%. According to Graton and Fraser (1935), more dense, minute rhombohedra1 colonies with local grain volumes of 74.1% constitute substantial portions, perhaps as much as one-half of the volume of generally less dense random packs, as indicated by inspection. The remainder %f the pack is represented by random colonies with grain volumes as low as those present in cubic packing (52%).

102

A. H. COOGAN AND R. W. MANUS

There are two reproducible states of random packing which have been called “loose” and “dense” by Scott (1960); the difference being about 3% grain volume. Scott filled rigid cylinders of known volume and nonrigid balloons with steel ball bearings 3.1 mm (very coarse sand) in diameter and by shaking cascaded them into containers. Graphical extrapolation provided a method of obtaining the grain volume of the‘packs in an infinite array, free from wall effects. He recommended the values of 63.7%grain volume for dense random packing and 60.1%grain volume for loose random packing of spheres. Following his work, Bernal and Mason (1960) investigated the coordination of close randomly-packed spheres by packing the same size steel balls and soaking them in black paint. After separation, the balls show points of contact not covered by paint. The mean number of contacts was 10-11 for a grain volume of 64%. This compares to the theoretical value of 8-10 contacts per grain for regular spherical packs (cases 2, 4, 5; Fig.3-5B, D, E, and Table 3-1) of comparable grain volume. They concluded that the volume of close random packing is repeatable, that it must be mathematically determinable, although so far undetermined, and that close random packing is one of minimum energy. Furthermore, any lowering of the stability of a packed bed must necessarily increase the distances between its spheres and increase the energy. Subsequently, Bernal and Finney (1967) further refined the estimates of volumes of random sphere packs. Using measured coordinates of a random model, which included a complete analysis of local density variations as well as determining an overall average in the absence of wall effects, they confirmed the values shown in Table 3-11 for spherical packs of loose and dense random packing.

Effect o f grain size Although theoretically grain size has no influence on the porosity or grain volume percentage of sediments, this does not prove to be true for assemblages of natural sands. As the sand size decreases, friction, adhesion TABLE 3-11 Loose and dense random packing of spheres T y p e of packing

Average grain volume (%)

Method of determination

Random, close

63.4

measurement of coordinates of sphere centers after removing outer layers

Random, loose

60.1

porosity measurements and extrapolation t o infinite volume without wall effects

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

103

and bridging become important because of the increasing ratio of surface area to volume and mass. Therefore, the smaller the grain size, the less the percentage of grain volume in randomly-packed sediments of small grain sizes. Determinations by Ellis and Lee (1919) on 36 samples ranging from coarse sand to silt show grain volumes (recalculated from porosity) varying from 46% to 61% (Table 3-111). The average grain volume of the 36 samples was 55.91%. As the grain size decreases, the tendency toward bridging and consequent looser packing is undoubtedly augmented by increased variation in the shape and differences in the method of deposition. No equivalent data are available for carbonate sediments.

Mixtures of sizes of grains The two main effects which occur when smaller spheres are added to a bed of larger spheres are (1) decrease in the grain volume percentage owing to forcing apart of the larger spheres, and (2) increase in the grain volume percentage because of filling of the voids, if they fit readily into the voids between the large spheres. The situation is considered here for binary, tertiary and more complex mixtures of grains.

Binary mixtures If an assemblage of uniform spheres is compared with an otherwise similar binary-size mixture, two effects are noticeable which appear to work in opposite senses but which are not of equal value. The result of size mixing is to increase the grain volume percentage by filling voids. As the large and small spheres become increasingly different in diameter, the increase in grain volume percentage is more pronounced. This applies both to the case of a single sphere in an assemblage of smaller ones and t o a number of large spheres in such an assemblage up to the point where the number of large spheres dominate the pack. Decreasing the grain volume percentage due to TABU 3-III Grain volume of sands sorted by size

Size grade*

Grain volume (%)

Coarse sand Medium sand Fine sand Fine sandy loam

61-59 59-52 56-51 5 0-4 6

*Fine sandy loam is a mixture of fine sand ( > .08 mm) and finer sediment where the fine sand content equals 50% o r more.

104

A. H. COOGAN AND R. W.MANUS

mixing spheres is caused by looser packing of small spheres about the larger ones, resulting in an increase of the “wall effect” throughout the pack. The ratio of the diameter of a small sphere, which can just pass through the pore throat between larger spheres into the interstitial void, t o the diameter of the larger sphere, is known as the critical ratio of entrance. This ratio is 0.154 for rhombohedral regular packing and is 0.414 for cubic packing (Fraser, 1935). When spheres of two sizes are mixed, the smaller size dominates the general structure of the pack so long as the proportion of these spheres is sufficiently great to keep most of the large spheres separated from one another. When the proportion of large spheres increases beyond this limit (small sphere control of the pack), two alternative situations arise depending on the relative sizes of the small and large spheres. When the small spheres have a diameter less than the critical ratio of entrance and the number of large spheres is sufficient so that they touch one another, are self-supporting, and bridge the smaller spheres, then the large spheres take control of the pack. From this point on, any further increase in the number of large spheres will decrease the grain volume percentage of the pack. On the other hand, when the small spheres have diameters exceeding the critical ratio of entrance, and the number of large spheres increases beyond the limit of domination by the small spheres, the two sizes will interfere with one another as more large spheres are added. This mutual interference will lower the grain volume percentage over what it would otherwise be. Nevertheless, grain volume percentage would still increase as the proportion of large spheres continues to increase. In the case of binary pack of sand and pebble-size grains, the same conclusions should be evident. Fraser (1935) tested the degree of disturbance of the packing of sand caused by a pebble contained within the sand. He experimented with four different sizes of pebbles collected from Revere Beach, Massachusetts, where the sand was fairly uniform and contained occasionalpebbles. At places where a pebble just protruded above the top of the sand, a pebblesand sample was carefully removed without disturbing the packing of the moist sand. The porosity of the entire sample was determined and recorded as total porosity. The volume of the pebble was determined separately and subtracted from the total volume (Table 3-IV). As expected, a large pebble has a strong effect on the grain volume percentage.

Multicomponent packs (Tertiary, Quaternary, Quinary) Measurements of the grain volume of various three-component mixtures of spheres have been made by Fraser (1935). He mixed sands in which the diameters of the spheres were approximately 1.3, 2.3, and 8.1 mm, so that each size was greater than the critical ratio for the next larger size in chance

105

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

packing. The lower grain volume percentages were invariably obtained in mixtures in which one size strongly predominated. With increasing complexity of the mixture, the grain volume percentage tended to increase. It is remarkable that average grain volume of 67% (33% porosity) can be secured consistently from very diverse mixtures of grain sizes (Table 3-V). One conclusion to be drawn from these data is that it apparently is impossible to predict grain volume percentage of a tertiary mixture of spheres on the basis of percentage of components. It is equally, if not more, unlikely that a useful prediction can be made for more complex mixtures. Much work remains to be done on the study of random sphere packings because there is no satisfactory geometrical probability analysis capable of explaining the value of the voidage of a randomly-packed bed (Ridgway and Tarbuck, 1967), and why it should be so very reproducible. We are even further from an explanation of the packing of spheres of different sizes or the packing of nonspherical particles. Experiments by Ridgway and Tarbuck (1967) indicate that the highest grain volume for multicomponent packs is achieved where the size ratio between the pairs of grain sizes is at least seven. Given a spherical grain with the radius r, the sizes of the secondary, tertiary and higher order sizes which will just fit in the void of a bed of rhombohedrally-packed primary spheres has been calculated (Table 3-VI). TABLE 3-IV Grain volume of beach sand around a pebble (From Frazer, 1935) ~~

~

Sample

Portion occupied solely by pebble (%)

Grain uolume of sample (%)

Grain uolume without pebble (%)

1 2 3 4

3.09 69.88 63.56 65.54

59.45 86.82 83.45 83.48

58.17 56.22 54.83 51.94

TABLE 3-V Grain volume of tertiary mixtures (calculated from measured porosity)

Sample

1 2 3 4

Percentage of grain sizes by sphere diameters 1.3 m m

2.3 m m

8.1 m m

10.51 29.57 41.43 10.17

9.63 41.39 29.84 44.48

79.86 29.04 28.55 45.34

Grain volume (76) 66.68 66.16 66.87 66.61

106

A. H. COOGAN AND R. W. MANUS

TABLE 3-VI Radii of spheres fitting in the void of rhombohedrally-packed primary spheres of a given radius r Sphere

Radius

Primary Secondary Tertiary Quaternary Quinary

r 0.414 r 0.225 r 0.177 r 0.116 r

Studies of natural sands of mixed grain sizes confirm some of the data obtained for spherical packs. Data from Steams (1927), strikingly portrayed in Fig.3-8, show that a wide range of mixtures of sand sizes may result in the same grain volume percentage. Similarly Gaither (1953) experimented with natural sands and showed that adding 8 g of coarse (0.5-1.0 mm) and medium (0.62-0.25 mm) sand t o 84 g of medium (0.25-0.5 mm) sand caused an increase of only 1.7% in grain volume. Earlier, King (1898, reported in Gaither, 1953) had determined that on decreasing the size sorting, by mixing sands having average diameters of 0.48 mm (medium sand) and 0.09 mm (fine sand), the grain volume of well-sorted, wholly fine sands increased from 60-65% to 76%. Packing he terogeneities Small-scale packing heterogeneities have been investigated by Morrow (1971) in terms of the distribution of pore sizes which is indicative of grain size heterogeneities in sediments without clay or cement. Packing heterogeneity occurs where there are variations in particle size or packing in different areas of a rock. For example, heterogeneities may be characterized by distinct local variations of particle sizes (Fig.3-9) and pore sizes. This could occur without variation in the average porosity or grain volume. In sands, such heterogeneities would be found where local variations in grain sorting exist. Another example of packing heterogeneity could be an essentially homogeneous random pack of spheres of equal size (Fig.3-10), but in which there are alternate regions of rhombohedra1 close packing and simple cubic packing. Interestingly, data provided by Brown and Hawksley (1945) can be plotted as a map (Fig.3-11) which shows that in a pack of uniform spheres the distribution of regions of tight, intermediate and loose packing is a random one. Capillary pressure drainage curves can be used to characterize packing heterogeneities in rock, because the slope of the curve reflects the pore size

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

107

Fig.3-8. Variations in grain size sorting of three sands with grain volumes between 67 and

68%. (Data from Stearns, 1927.)

Fig. 3-9.Heterogeneous random distribution of primary and secondary size spheres, shown as a monolayer. Radius of secondary spheres is less than 0.4 r and greater than 0.1 r of primary sphere (radius = r), or smaller than the critical value for a cubic pack. This fabric could result from burrowing by organisms in lagoonal sands.

108

A. H. COOGAN AND R. W. MANUS

Fig.3-10. Heterogeneous packing of uniform spheres containing central portion of rhombohedra1 packs R and marginal cubic packs C, shown as a monolayer. Large pores are present at transitions from one packing to the other.

Fig.3-11. Distribution of density of packing in tightly-packed groups of uniform spheres. Percentage for (1) tight packing (grain volume > 79%) is 44% - light dot pattern; (2) intermediate grain volume (79%-7'1%) is 17% -dark dot pattern; and (3) loose grain packing (grain volume < 71%) is 40% - white. (Based on data from Brown and Hawksley, 1945.)

distribution of porous sand. In addition, larger interconnected pores tend to drain faster than smaller ones. The capillary pressure curve, according to Morrow (1971),proved to be useful in determining the apparent pore size (hence, grain size) distributions of sands. Based on experiments with artificial packs in which clusters of particles are surrounded by a matrix of

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

109

coarser particles, Morrow concluded that the irreducible wetting phase saturation of such mixtures is much higher than the saturations of corresponding well-mixed, more homogeneous aggregates of the same particles. Because the heterogeneity of packing largely determines the magnitude of the irreducible wetting phase saturation (Fig.3-12) in completely wetted system, the capillary pressure curve in turn provides an index of packing heterogeneity. Accordingly, a parallel bundle of capillaries, or a rock with any other pore configuration that allows complete drainage, has zero heterogeneity. Theoretically, rhombohedral and cubic packs having spheres of equal size will have irreducible saturations of about 5.5% and 4.276, respectively. The liquid will be held as pendular rings at points of contact between particles (Harris and Morrow, 1964). In random packs of equal spheres, irreducible saturation falls in the range of 6-876, representing only a slight loss of homogeneity from the regular packs. Other well-mixed packs, including artificial sands (Morrow, 1971) give values in the range of 6-1096. The irreducible saturation values of 15-4076 are commonly observed for sedimentary rocks and indicate a significant variation in packing heterogeneity and possibly cementation. The effects of porosity and permeability on irreducible wetting phase saturation are shown by Chilingar et al. (1972) for a variety of carbonate rocks. Obviously a variety of packing types can give the same irreducible saturation. The reasons for a given region remaining more saturated than other areas might range from gross differences in particle size to subtle differences in packing arrangement. Also, even though packing of coarse

- - - - - - - - - - - - ,

I

110

A. H. COOGAN AND R. W. MANUS

particle clusters surrounded by a finer matrix are clearly heterogeneous, no indication of this may be given by the capillary pressure curve, because the packing will appear to be homogeneous and the sample will drain to a uniform saturation throughout on a gross scale. Nevertheless, natural sands will likely contain mixtures of various types of heterogeneity (Graton and Fraser, 1935) so that, on the average, values for irreducible wetting phase saturation should provide a reasonable measure of rock heterogeneity where cementation is insignificant. Where capillary pressure curves, porosity, and permeability are obtained for rocks which have also been examined in thin section for patterns of grain distribution, pore types and cement, these physical measurements will enhance understanding of packing in rocks and hence the state of compaction (see Chilingar e t al., 1972). Recent and Pleistocene oolitic grainstones from the Bahamas and Florida were characterized by Robinson (1967) in terms of porosity, permeability, and cgpillary pressure curves. Slightly cemented, slightly altered, well-sorted oolitic grainstone with 41.5% porosity and 8,895 md permeability had a low threshold entry pressure of 0.2 kg/cmZ and almost no displacement zone. The Gshaped capillary pressure curve was typical of well-sorted, well-connected pores. In contrast, altered, leached and cemented oolitic grainstone, with equally high porosity (45.7%), lower permeability (935md) and less well-sorted pore types (mixture of intra- and interparticulate pores), had lower threshold entry pressures, but a wide displacement zone of nearly 7.0 kg/cmZ. The irreducible wetting phase saturation varied as much as 30% for these oolitic grainstones which had little difference in porosity.

Random packing o f irregular grains The shape of sedimentary grains is probably never spherical, not even in oolitic sands, and these irregularities in shape should result in a larger possible range of grain volume percentages in sediments, because irregular grains theoretically may be packed either more tightly or more loosely than regularly-packed spheres (see Chapter 2). It is difficult to determine the effect of grain shape on grain volume percentage in natural sediments because of the difficulty of obtaining odd-shaped particles of the same size by screening. Determination of the effect of grain shape on porosity and packing is commonly hindered by the second independent variable, grain size. As a result, a study of the average packing conditions by determining the porosity in a sand, which has grains of different sizes and shapes, generally supplies little information concerning the effect of shape on grain volume. Studies of carefully-sized materials, which range in shape from spheres to flat plates, by Fraser (1935), however, showed expectable

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

111

decreases in grain volume percentage with departure from spherical shape (Table 3-VII). For lead and sulfur shot, the values at loosest packing are not reliably comparable with those for natural sands because the greater specific gravity of these materials affects the compactness of the unjarred assemblages. The shot was screened and the portion used had an average diameter of 1.5 mm (coarse sand). The porosity was measured before and after the shot was consolidated by jarring when dry and when saturated with water. If the averages of the lead and sulfur shot tamped down are taken as representative of uniformly-sized spherical grains, then packs of dry spheres have a grain volume of 53.28% and wet ones 66.68% (Table 3-VII). Comparing the other samples with the shot, two have higher grain volume percentages and two lower. The shape of the marine sand grains is reported to be fairly uniformly disk-shaped and may be packed more tightly to the third dimension, with flat sides together, which would allow for the high grain volume percentage displayed. The graih volume percentage of beach sand does not differ widely from that of the spherical shot. Assuming that variations in grain volume of compacted dry sands are due to the influence of the grain shape, there exists a range of 9.63%grain volume variation (from 66.22 t o 56.59%)for a suite of rather typical samples. Wet meterials pack more loosely than dry materials owing to the bouyancy of water. The difference between packing of rounded and angular grains (wet and dry) in compacted samples, however, proves that wetting the sand increases the effect of angularity on the packing. Of the samples tested, flat and needle-like grains have the greatest effect on packing. For example, TABLE 3-VII Influence of grain shape on grain volume of wet and dry, loose and tamped sands (After Fraser, 1935) Material

Specific gravity

Grain volume (%) wet

dry

Lead shot Sulfur shot Marine sand Beach sand Dune sand Crushed calcite Crushed halite Crushed mica Crushed quartz

11.21 2.02 2.68 2.65 2.68 2.66 2.18 2.83 2.65

loose

compacted

loose

compacted

59.94 56.62 61.48 58.83 58.83 49.50 47.95 6.47 58.80

66.72 66.65 63.22 63.45 66.40 59.24 56.49 13.38 58.80

57.60 55.86 57.04 53.45 55.07 55.50 7.62 56.12

61.11 61.76 64.96 61.54 60.66 57.26 12.72 56.04

112

A. H. COOGAN AND R. W. MANUS

crushed mica had a grain volume of less than 10%.This amazingly low grain volume was not increased above 14% by prolonged jarring. Pressures sufficient t o burst a strong glass container were required to increase the grain volume percentage of the wet mica to 32.6%(Fraser, 1935). Angularity, or departure from spherical shape, may increase or decrease the grain volume of packs. Most commonly it decreases the percentage grain volume and, according t o Fraser (1935), the only type of angularity found to increase grain volume is that in which the grains are mildly and uniformly disk-shaped. Changes in grain packing with jarring It has been shown that any aggregation of grains has a fairly definite range of grain volume which represents varying approaches to perfection in packing. Moderately well-sorted, medium-grained, randomly-packed natural sands have grain volumes between 53.5% and 62.2% depending on the looseness of packing, angularity of the grains, wet or dry state when measured, and other factors. Fraser (1935) took the analysis one step further by tamping a beach sand with a grain volume of 54.0% (wet) and 53.70% (dry) for 12 min. The grain volume increased nearly 10%. Further jarring produced no additional settling, however. In fact, .further jarring loosened the packing (Table 3-VIII). Pebble and sand grains are deposited in a more stable state than are smaller particles, because the effect of buoyancy by the water during sedimentation is smaller for coarser grains. Accordingly, the grain volume percentage and packing configuration of sands and gravels as initially deposited cannot be greatly increased until stress is sufficient t o crush the grains, or solution removes portions of the grains to provide space for closer packing. When artificially deposited under water, sand derived from the St. Peter Sandstone (Athy, 1930) could be made to settle about 11%of initial TABLE 3-VIII Grain volume variations with duration of tamping in beach sand (After Frazer, 1 9 3 5 ) Tamping duration (min)

Grain volume (%)

0 1 2 4 8 12

54.00 57.50 59.23 61.53 61.74 61.74

53.70 60.47 62.00 61.80 62.18 62.18

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

113

volume by continued jarring at normal atmospheric pressure. When a pressure of 280 kg/cmZ was applied, the additional increase in compaction was only 2%.

Summary of packing of natural sands Natural, randomly-packed, nonspherical, moderately well-sorted and rounded sands have average grain volumes between 50 and 70%; the majority of values falls between 54 and 67%. Based on these data, Rittenhouse (1971) concluded that regular orthorhombic packs (grain volume of 60.46%) of spherical particles approximate natural randomly-packed sands on the average, but that the analogy is semi-quantitative. In measuring oolitic grain volume in thin section, Coogan (1970) found that undisturbed, in-place, oolitic sand has an average grain volume of 64.8% and subaerially weakly cemented oolitic grainstone of Pleistocene age has a grain volume of 64.4%. Experiments have repeatedly shown (Gaither, 1953) that well-sorted, wellpacked, wet sands have grain volumes about 67%. There seems to be overwhelming evidence that the values cited are typical of a wide range of natural sands of varying composition and sorting. Somewhat lower values, however, were obtained by Pryor (1971) from porosity data on modern sands collected from varied environments. Expressed as grain volume percentage, Pryor found values between 50% and 60% (average 55%)for modern river sands, 44-61% (average 54%) for beach sands, and 49-56s (average 52%) for dune sands. His values appear to be for angular sands and there may be some disturbance from side-wall effects when the samples were pushed into small-diameter core barrels. At present, the point of zero compaction for common, randomlyoriented, mixed sands of moderate sorting, reasonably high sphericity, and moderately close packing can be taken at about 6 0 - 6 5 % grain volume. For further refinement in specific cases, artificial mixtures must be prepared and tests made t o determine additional reference points. This is especially true for Recent carbonate sediments and for ancient carbonate rocks. EVIDENCE FOR COMPACTION OF CARBONATE SANDS

General statement Aside from the nearly universal recognition of stylolites in limestones, most references t o compaction of carbonate sands refer in a general way to apparent “overclose” packing, t o fracturing of grains, or to grain contacts which appear to reflect solution or impression. All of these effects also are

114

A. H.COOGAN AND R. W. MANUS

observed in barely compacted carbonate sands (Carozzi, 1961; Bishop, 1968; Purdy, 1968). Of course, many do result from compaction but these features do not in themselves provide any direct lead to the amount of compaction which has occurred in the rock since deposition. Few attempts have been made t o quantify compaction, that is, to measure the amount of compaction which has occurred in carbonate grainstone (Kahle, 1966; Coogan, 1970), owing to the difficulty in determining a value for zero compaction. The present state of the art recognizes two useful approaches t o measuring compaction in lime grainstones and, together with well-established methods of carbonate petrography, these can provide considerable information on the degree of compaction. Measuring packing and compaction, petrographic techniques - packing density and packing index The two principal measurable effects of the compaction of carbonate sand are: (1)increase in grain volume percentage as compared to the original uncompacted sediment, and (2) increase in closeness of the grains as reflected in the increased number of grains which touch each other in a compacted sand. The first effect, increase in grain volume percentage, can be directly measured in thin sections of lime grainstones using one or more of the point counting methods. In 1956, Kahn proposed a new parameter, Packing Density and defined it as:

i 1 n

PD

=

m x gi/t X 100 i=l

(3-1)

where n is the total number of grains in a given traverse across the thin section, gi is the grain intercept of the ith grain, m is the magnification constant, and t is the length of the traverse. Packing Density is essentially the equivalent of grain volume percentage as measured in two dimensions of a thin section and is closely comparable with three-dimensional grain volume if sufficient area of the thin section is traversed. A comparison of grain volume percentage determined by weight and volume measurements and by point counting (Gaither, 1953) show that the average of about,7% greater grain Packing Density was obtained from thin section measurements. In measuring oolitic sand volume of Bahamian samples, a grain volume calculated from volumetric measurements was about 2% lower than that determined using thin section measurements. Most of the difference is attributed by Gaither (1953) to shortcomings of preparation techniques. A simpler form of the Packing Density expression can be used in grainstones where no consideration is given to recording cement and pore space

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

115

separately. A simplified expression used in measuring compacted oolitic grainstone is (Coogan, 1970): -

xPD%

=

(gi/TL) x 100

(3-2)

where XPD is the arithmetic mean of the separate measurements made on all traverses of the thin section and TL is the total of all traverses. This method of determining Packing Density is shown in Fig.3-13, where a single traverse line crosses six ooliths, pores, cement and some shaly matrix. The six ooliths occupy 73% of the linear distance of the traverse. The summation of successive traverses yields the mean Packing Density of the sample. In this example, the volumes of cement, pores, and matrix are 7, 12, and 896, respectively. As compaction pro.ceeds, grains are pushed closer together resulting in a higher grain volume percentage, cement, and matrix volume, and lower pore volume. The minimum value for Sacking Density corresponds to the minimum value of the grain volume percentage for the uncompacted sediment; the maximum value is 100%. The Packing Density measurement is itself affected by compaction. Calculation of the standard deviation for several samples of uncompacted oolitic grainstones shows values as high as 10% for one standard deviation (PD = 65% f 10%). Furthermore, the short traverse data (Fig.3-13) show a nearly normal distribution of Packing Density values, with a mean at about 65% (Fig.3-14) and a small value for kurtosis. In comparison, data from deeply-buried Jurassic Smackover oolitic grainstones (Fig.3-14) show that 0

20

SCALE -100 UNITS

40

60

80 1

1

100 ,

Fig.3-13. Illustration of the measureTent of Packing Density and Packing Index using a thin section of spheres, shown as a monolayer. (a) Number of ooliths crossed by traverse, N = 6 ; (b) sum of grain intercepts, g i : 1 0 + 1 5 + 14 + 13 + 1 5 + 6 = 73; (c) total length of traverse, TL = 100 line units; (d) packing density calculation for single traverse, PD% = (gi/TL)x 100 = 73%; (e) grain-to-grain contacts, GC = 2; ( f ) packing index calculation for single traverse, PI% = ( G C / N ) X 100 = 2/6 X 100 = 33.3%; (g) other data: grains touching = 3; grain-to-matrix contacts = 1 ; grain-tocement contacts = 3; grain-to-pore contacts = 2; length of matrix intercepts = 8 units; length of cement intercepts = 7 units; length of pores intercepts = 1 2 units. Stippled pattern is cement; white is pore space; dashed pattern is shaly matrix.

116

-

A. H. COOGAN AND R . W. MANUS

50

so

f

$

40

c

4oji

30

.2 30

20

20

20

10

10

10

. 2 -

30

80 60 40 PD, K

A

PD,K

B

P D, K C

Fig.3-14. Histograms of Packing Density sorting for three oolitic grainstones. A. Uncompacted oolitic sand from the Bahamas (XPD = 64.4%; SD = 10.1%). B. Jurassic Smackover grainstone from a depth of 3,150 m (2PD = 90.5%, S D = 5.8%). C. Jurassic Smackover oolitic grainstone from a depth of 3,687 m (2PD = 8 3 . 1 % , S D= 7.5%). Width of histogram columns is equal t o 10%Packing Density interval. PD = Packing Density, f = frequency, X = mean.

the mean shifts toward the 100%limit with strong.asymmetry and a forced reduction in the value of the standard deviation. Comparing the three samples illustrated in Fig.3-14, it can be seen that with increasing compaction, the mean of the Packing Density shifts, the standard deviation is somewhat reduced, and a strong kurtosis may develop. The reason for the variation in kurtosis is not clarified solely by the histograms. Inspection of the samples, however, shows that grain-size sorting is a crucial factor. In the case of the Smackover oolite buried to 3,687 m the grains are poorly sorted. The tendency of large grains to bridge across the smaller ones would cause the smaller ones t o absorb less of the compactive stress. This could result in certain areas of the rock exhibiting a lower degree of compaction than others. In the case of the sample from a depth of 3,150 m, the grain size sorting is good and, hence, the stress should be more evenly distributed. The second feature, increased number of grains in contact, is representedin thePuckinglndex proposed by Masson (1951)and used with modification by Kahle (1966) and Coogan (1970) in measuring the compaction of oolitic and pelletoidal lime grainstones. As originally defined, the Packing Index is based on grain boundary measuremen&. Use of Masson’s index requires the counting of intersections (one per grain) between traverse lines and grain boundaries (either grain-to-grain, grain-to-cement, or grain-to-pore space). Contacts are counted as the traverse leaves the grain. Masson’s formula is: PI = 100 (Ng/Ng + N n )

(3-3)

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

117

where Ng is the number of grain-to-grain contacts and Nn is the number of grain-to-cement contacts. In 1956, Kahn proposed a similar measure of closeness of grains called the Packing Proximity for which the formula is: Pp

=

q / n X 100

(3-4)

where q is the number of grain-to-grain contacts, n is the total number of contacts (as well as the total number of grains), and O < q < n. The maximum value for this expression is 100%. For oolitic grainstones, the expression may be written as:

-

x PI%

=

GC/N X 100

(3-5)

where GC is the number of grain-to-grain contacts along the line of traverse, and N is the number of grains along the same traverse. Values for repeated traverses are summed to derive the mean Packing Index. An illustration (Fig.3-13) of the method of measurement of Packing Index and Packing Proximity shows that Masson’s (1951) Packing Index (in the example, 40%) is not the same as Kahn’s (1956) Packing Proximity (in the example, 53.3%). The latter index is the same as the simplified Packing Index used by Coogan (1970). The rationale behind the use of the Packing Index is the idea that lime grainstones which are compacted should have more grains touching each other than those which are uncompacted. All packing indices measure similar relationships. The Packing Index is more quickly determined than the Packing Density. It is less useful in an evaluation of compaction in rock, however, because the Packing Index is affected more by the total number of grains counted, the shape of the grains, their sorting, and the effects of cementation and diagenesis which change grain contact configurations. The dependency of the Packing Index on the total number of grains counted may cause difficulties as can be seen in Fig.3-15 and Table 3-IX. In Fig.3-15, 10 spherical grains are drawn to occupy a traverse length of 100 linear units through the centers of the spheres in row 1 for which N = 10, TL = 100, PD = loo%, GC = 9, and PI = 90% (Table 3-IX). In this situation the maximum Packing Index for the line is 90% and not 100% as stated by Kahn (1956), because,there are only 9 grain-to-grain contacts possible (contacts are counted as the traverse leaves the grain). In rows 2-6 (Fig.3-15), 5 spheres occupy 50% of the traverse length but are arranged so that all the grains touch each other in row 2, whereas none touch in row 6. Rows 3-5 have intermediate numbers of grains touching each other. As shown in Table 3-IX, the Packing Density for all the spheres in rows 2-6 is the same (50%), but the Packing Index ranges from zero to 80% depending on the spacing of the grains. Approximately

118

A. H. COOGAN AND R. W. MANUS

100 L I N E

4

UNITS

Fig.3-15. Diagram used in measuring Packing Density and Packing Index for single-line traverses of 5 and 10 uniform spheres in a row, as presented in Table 3-IX.

TABLE 3-IX Relationship among the Number of grains (N), Grain Contacts ( G C ) , Grains Touching ( G T ) ,Packing Density (PD) and Packing Index (PI) for idealized packed spheres (Fig.3-15); number of grains and grain contacts were measured in thin section N

TL

xgi*

PD%

GC

GT

PI%

Fig.3-15, row no.

10 9

100

100

100 90 90 80 80 80 50

9 8 7 7 6

10 9 8 8 7 6

90 89 78 87

-

62 80 60 40 20 0

-

9 8 8 8

5 5 5 5 5

100 100 100 100 100 100 100 100 100

100

90 90 80 80 80

50 50

50

50 50

50 50 50 50

5

4 3

2 1 0

5

4

3

2

0

75

* c g i = sum of measurements of grain intercepts of the ith grain.

I

2

3 4 5 6

119

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

200 or more grains must be counted before a maximum value for the Packing Index is approached for any given Packing Density (Fig.3-16). Wide ranges in grain size, as well as differences in apparent size owing to the random intersection of the thin section plane with the grain, also reduce the chance of intersecting grain-to-grain contacts and affect the Packing Index measurement. In spite of the difficulties of directly relating Packing Density for short traverse data t o Packing Index (Table 3-IX), it does appear possible to relate the mean Packing Density for a number of samples t o the mean Packing Index. In Fig.3-17, the mean values of the Packing Index are plotted versus mean Packing Density for short line .traverse measurements on 24 thin sections of oolitic grainstones. A linear regression was calculated using the expression: y (PI%) = a

+ b x(PD%)

(3-6)

'

where (X over i = 1 . . . n ) a = Xyi - bCxi/n and b = [ n C x i ~ i_ Cxiyi]/ [nCxf - ( C x i ) ' ] (Freund and Williams, 1958). Thus, x(Pl%f =

0

100

200

300

400

500

600

100

8bO

900

1000

GC/N

Fig.3-16. Relationship between Packing Index (PI)and G C / N ratio for different numbers of grains. The maximum PI increases rapidly from zero with increasing numbers of grain contacts per number of grains. Diagonal lines show plot of G C / N for samples of differing total number of grains (e.g., N = ZOO). As N a n d GC/N increase above 100, the maximum PI flattens. PI' = Packing Index, N = number of grains, GC = number of grain contacts, Max. = maximum

120

0

A. H. COOGAN AND R. W. MANUS

50 PD, %

100

Fig.3-17. Relationship between the mean Packing Index ( P I ) and mean Packing Density

(PD) based o n short-line traverse measurements on thin sections of compacted and

uncompacted oolitic grainstones. Solid points are individual line measurements. Small and open circles are mean values for 40 traverses each.

X

's

-125 + 2.13?(PD%). This result is considered tentative because few samples counted were completely uncompacted grainstones. Nevertheless, the regression is based on traverses of over 8,000 grains in oolitic grainstones from surface and subsurface localities ranging in age from Mississippian to Recent. The minimum value for the mean Packing Index varies with the shape of the grains and the grain volume. Randomly and loosely arranged packs of spherical grains have values as low as 2.5%, but the normal range for Recent oolitic sands is about 18-20%; the maximum is 100%. The minimum Packing Density projected for zero Packing Index is 54% (Fig.3-17), a value close to that of the grain volume percentage of the ideal cubic (case 1) packing (Fig.3-5A). In sunimary, these two measurements of packing are important in establishing the extent of compaction. The Packing Index is less useful but more rapid to calculate. It may disclose one anomalous compaction feature, the one indicated by high Packing Density and low Packing Index, involving

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

121

a situation in which the grain volume percentage has been reduced but the grains seldom touch each other. In general, use of the Packing Density is recommended for routine thin-section analysis of compaction.

Other measurements and evidence of compaction Many petrographic features may indicate that carbonate particles have been subjected t o compactive stress. Special notice must be taken of the work on grain-contact relationships in clastic sands. The change from the apparent “floating” grains and grains with tangential contacts, to grains having concavo-convex and sutured microstylolitic contacts (Fig.3-18) is evidence of compaction. In her study of pore space reduction, Taylor (1950) considered the number of impressed and stylolitic contacts which were an index of compaction. This procedure is risky, however, because the original shape of the grain (spherical, elongate, discoidal, or platy) will determine to a great extent the shape of the grain contact. For example, longitudinal contacts (Fig.3-18) occur readily between elongate grains as a result of sedimentary packing without any influence of compactive stress. The composition, hardness, ductility, and susceptibility to solution of grains are important in determining the final grain-contact shape as a result of compaction. Hard quartz grains tend to penetrate soft calcite ooliths without producing microstylolitization, whereas adjacent ooliths in the same thin section show mutual solution features. In a study by Hays (1951, cited by Gaither, 1953) it was noted that quartzose sands with a large amount of carbonate grains had an abnormally high percentage of sutured contacts compared with other quartzose sands buried to equivalent depths, but devoid of carbonate grains.

Fig.3-18. Petrographic effects in randomly-packed, compacted oolitic grainstone, shown as a monolayer. A = pore space, B = tangential contact between grains, C = cement, D = microstylolitic contact between grains, E = ruptured cortex of oolith, F = calcite-filled, prelithification fracture, G = longitudinal contact, H = oolith, I = authigenic quartz, J = indented grain contact, K = authigenic quartz incorporating microstylolite, L = macrostylolite, M = postlithification fracture, and Q = polycrystalline quartz grain indenting ooliths

122

A. H. COOGAN AND R. W. MANUS

The data from Taylor (1950) for sandstones show a definite trend toward increased number of contacts per grain and more pressure solution with increasing depth (Table 3-X). There is a corresponding increase in values of Packing Density and Packing Index, because the counting of the total number and/or the number of different kinds of grain contacts is a form of packing index measurement. She reported between 0.63 and 1.6 contacts per grain for randomly-packed artificial sands and Gaither (1953) estimated a value of 0.85 for the number of contacts per grain for freshly-deposited clastic sand. Naturally, the size, shape and sorting of the grains are important factors in determining the number of grain contacts. The similarity between the determination of Packing Index and Taylor’s method of counting types of grain contacts suggests that the latter has little to add to the measurement of compaction experienced by a sand. The study of the kinds of grain contacts, the mineralogical composition of the grains, and their relationships to cement, however, may yield useful information concerning compaction of sands. In a comparison of artificially-compacted, coarse-grained carbonate sediments with naturally-compacted grainstones, Ayer (1971) recognized 12 structural and petrographic deformation features related t o individual grains. Somewhat modified, these are: TABLE 3-X Number and types of grain contacts in synthetic and natural, buried, Wyoming sands (From Taylor, 1 9 5 0 ; Gaither, 1 9 5 3 ) Grain contact data

Synthetic sands

Formation name and depth

Taylor

Mesa Verde, 865 m

Average 1.6 number of grain contacts Floating 16.6 grains (%) 59.4 Tangential contacts (%) Long 40.8 contacts ( a ) Concavo0 convex (%) Micro0. stylolitic (%)

Gaither

0.85

Shannon, 1,370 m

Lower First Wall Creek, 2,075 m

Morrison, 2,540 m

2.5

3.5

4.4

5.2

46.0

0.3

0

0

0

77.0

51.9

21.4

0.9

0

17.0

38.1

59.8

51.6

45.0

6.0

9.6

19.1

28.5

23.1

0

0

0

18.5

31.8

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

123

(1) Point contacts with no grain interpenetration (also called tangential contacts). (2) Concavo-convex contacts, which result from the yielding of one grain. (3)Linear contacts which result from mutual yielding of grains; these are not t o be confused with point contacts between elongate grains. (4) Sutured contacts (microstylolitic), which show small grain interpenetration. ( 5 ) Plastic flow or distortion of grains. (6) Crushing, which results from repeated microfaulting or shattering. (7) Radial fractures which are confined to the radius of the grain and do not pass through the center. (8) Diagonal fractures of the grain from the surface through the center. (9) Faulting or rupture, which breaks the grain in two or more parts that remain nearly contiguous. (10) Splitting, which breaks and separates the grain into two more-orless equal parts; the displacement is perpendicular to the fracture trace. (11) Buckling (crinkling) of the outermost portion of the grain, which results in a small sharp crest. This feature is common to some oolitic grainstones. (12) Spalling, which flakes off the outermost portion of the grain. This feature is common to some deformed oolitic grainstones. It is not clear whether or not concavo-convex contacts are distinctly separate from flow features. It appears that several of the 12 deformational types are related not only to the degree of compaction but also to the spherical shape of the oolith grain. Such deformational features should be rare or indiscernible in non-oolitic carbonate particle grainstones. Another approach to measuring compaction in thin section using graincontact relationships was proposed by Allen (1962). He measured the degree of indentation of grains by determining the ratio of the length of the grain margin touching another grain (fixed margin) to the length of the remaining boundary (free margin), which is in contact with pore space or cement. An indented grain is considered as fixed when the fixed margin exceeds the free margin in length and as free in the opposite case. His measurement of indentation is called the Condensation Index: Condensation Index = % Fixed Grains/% Free Grains

(3-7)

In part, this index relates the number of “floating” and tangential grains to the number of impressed and sutured grains and is similar to Taylor’s approach. A further attempt to use grain contacts in the form of the free and fixed margin ratio develops an index of compaction defined as:

124

A. H. COOGAN AND R. W. MANUS

Fixed-Grain Compaction Index = (Number of Fixed Grains/Total (3-8) Grains) X 100 In the latter case, the Fixed Grain is one with more than 50% fixed margin (Allen, 1962). The maximum possible value for the Condensation Index is infinity, whereas the maximum value for the Fixed-Grain Compaction Index is 100%; the minimum value for both is zero. A comparison of these indices with Packing Density, Packing Index, average number of grain contacts per grain, and Compaction Index is shown in Table 3-XI for four oolitic grainstones (Fig.3-19-3-22). For example, the Condensation Index of the Bahamian uncemented oolitic sand is zero and that of the strongly compacted Smackover oolite is 0.62. The latter is an inconveniently lowsounding value for a rock with a Packing Density of 95.9%, considering the fact that the maximum value for the Condensation Index is 99%to infinity. The Condensation Index scale is not an arithmetic scale and the values do not progress uniformly with increasing compaction. Values for the FixedGrain Compaction Index are zero for the uncompacted Bahamian oolite and 38% for the Smackover oolite. On the Compaction Index scale derived from the Packing Density (Coogan, 1970), to be discussed later, Smackover oolite is 87% compacted. Unfortunately, the indices used by Taylor and Allen have not been related to the uncompacted state of carbonate sands so as to allow an estimate of the zero point of compaction on their scales. In addition, it is TABLE 3-XI Comparison of various measurements of compaction for four oolitic grainstones (Fig.3-19-3-22, Table 3-XV) Measurement

Mean Packing Density (%) Mean Packing Index (%) Compaction Index (%) (Coogan, 1970) Condensation Index (%) (Allen, 1 9 6 2 ) Fixed-grain Compaction Index (%) Grain contacts per grain Figure no.

Specimen*

Maximum possible value

1

2

13

3

64.8 18.6 zero

70.3 22.5 15

82.9 18.6 51

95.9 66.5 87

zero

0.13

zero

10.8

1.2 (3-19)

1.4 (3-20)

*For sample descriptions, see Table 3-XV. **From Marvin (1939).

0.11 8.1 0.8 (3-21)

0.62 38 3.7 (3-22)

100 100 100 99 to infinity 100

14.1**

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

125

Fig.3-19. Loose, uncompacted, oolitic sand, impregnated with plastic in situ on an oolite bar top to preserve fabric. Browns Cay, Bahamas.

Fig.3-50. Weakly-compacted oolitic doloarenite, a dolomitized oolitic grainstone. Jurassic Smackover Formation, Pan American Petroleum Corp., No. 1, Parker, Van Zandt Co., Texas, depth = 3,917 m.

doubtful that their scales could be used with much precision for irregularly shaped grains of bryozoans, crinoid columnals, plate algae or other biologically derived carbonate particles. For example, the Condensation

126

A. H. COOGAN AND R. W. MANUS

Fig.3-21. Compacted oolitic grainstone, Mississippian Ste. Genevieve Formation, Roane Co., Tennessee, outcrop.

Fig. 3-22. Compacted oolitic grainstone showing interpenetrating grains, microstylolitic contacts and cement. Jurassic Smackover Formation, Tenneco Oil Co., No. 1, Lowe, Clairborne Parish, La., depth = 3,149 m.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

127

Index is zero for Bahamian oolite (Fig.3-3A) and gastropod sand (Fig.3-3B), yet the Packing Index of the two sands is quite different. Inasmuch as both of the indices using the fixed-grain concept are based on grain-contact relationships, both suffer from the same deficiencies in use that one encounters on using the Packing Index. These indices, therefore, cannot be recommended as measurements of compaction. Compaction index A Compaction Index based on Packing Density was formulated by Coogan (1970) for oolitic grainstones. An increase in Packing Density reflects the movement of grains closer together, presumably as the result of increased overburden pressure. Thus, any increase in Packing Density may be thought of as an increase in the amount of compaction from some Packing Density value, which represents zero compaction for the particular carbonate sand (depending on the particle shapes, sorting, and sedimentary packing configurations), to 100% Packing Density, representing a rock composed entirely of grains or one that is 100%compacted (no porosity). It is recognized that subsequent compaction through gross dissolution of calcium carbonate along stylolitic surfaces might occur (Park and Schot, 1968). Based on the concept of increasing Packing Density, a revised Compaction Index is presented (Fig.3-23), which arbitrarily relates Packing Density or grain volume percentage in thin section to the amount of compaction. New data have been added for special carbonate grain shapes. An initial zero compaction index value was calculated for several different kinds of carbonate sands (curves A-F, Fig.3-23). For sands composed principally of ooliths, the value of 65% Packing Density or grain volume percentage was taken as zero compaction based on determinations of the average Packing Density of unburied, naturally-packed, well-sorted oolitic sand from the Bahamas (Coogan, 1970). Two different kinds of samples were examined. One, a Recent oolitic sand from an oolite bar top, 1.5 km east of Browns Cay was sampled by impregnation with epoxy resin, preserving the undisturbed cross-bedding fabric of the sand. Thin sections of this sand had a Packing Density of 64.8%, a sample standard deviation of lo%, and a Packing Index of 18.6%.A check on these values was made by measuring the Packing Density of oolitic grainstone from subaerially weakly-cemented oolitic rock of Pleistocene age on Joulters Cay, northeastern edge of Andros Island, Bahamas. The Joulter’s Cay oolitic grainstone had a Packing Density of 64.4%(SD = 9.6%)and a Packing Index of 21.6%. As a point of reference, the mean grain volume of 9 cemented Pleistocene oolitic grainstones from Florida and the Bahamas, converted from Robinson’s (1967) bulk porosity data, is 60.6%(39.4%porosity; 6,777 md permeability). Six very lightly cemented oolitic grainstones had a mean grain

128

A. H. COOGAN AND R. W. MANUS PORE AND CEMENT VOLUME, K

00

CI, %

100

80

60

40

20

1

QRAIN VOLUME, %

Fig. 3-23. Relationship between Compaction Index ( C I ) and grain volume percentage (hence also pore plus cement volume) for selected mud-free carbonate sands. The range of grain volume percentage is from 10 to 100%;the range of pores plus cement is from 90 to zero % depending on the original packing (curves A -F ) and, amount of compaction. A = oolitic grainstone, B = grapestone, C = angular, mixed skeletal sand, D = valves of the pelecypod Noetia ponderosa Say, E = valves of Anomia simplex Orbigny, and F = red algal particulate sand (data from Dunham, 1962).

volume of 58.7% (41.3% porosity; 47,000 md permeability). Based on earlier cited values of grain volume of randomly-packed spheres and the Packing Density measurements from the oolitic sands, a Packing Density value or grain volume percentage of 65% was assumed to be an average value for the state of zero compaction for oolitic sands on the Compaction Index scale. Attempts to relate other types of carbonate grain sands to this scheme are just beginning, because of the lack of knowledge on the average packing of most other typical carbonate sands. Nevertheless, new determinations by the writers on shell materials collected from the west coast of Florida and from the Bahamas, backed with data from the literature, provide useful approximations of the grain volume of grapestone and single-component and mixed shell sands. These data provide zero compaction reference points for the Compaction Index chart (Fig.3-23).. The data for grapestone provided by Fruth et al. (1966) were checked by determining the porosity of grapestone sand collected in the Berry Islands, Bahamas. The most densely packed samples with the least wall effect have a grain volume of 55.396, about 3% above the value determined by Fruth et al. (1966). The 55.3% value is used as the zero compaction point on the Compaction Index line for grapestone (curve B, Fig.3-23).

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

129

Estimates for skeletal lime grainstones (curves C-E, Fig.3-23) are derived from three sets of data for different-shaped particles. A shell “hash” was collected from a 10-cm high beach ridge on the causeway between Fort Myers Beach and Sanibel Island, Florida. The mean grain volume based on porosity determinations for this coarse sand is 39.5%. The grains are very angular and consist of broken particles of numerous bay or inlet dwelling pelecypods and gastropods. The sample has a mineralogical composition of about half calcite, half aragonite. The zero value for mixed skeletal grainstone Compaction Index line (curve C, Fig.3-23) is based on analysis of this collection. In determining the porosity for more loosely packed skeletal sand, a collection of valves of the taxodont pelecypod Noetia ponderosa Say, the ponderous ark shell, were used. The sample was scooped from piles of shells washed up on a narrow beach facing the Gulf of Mexico, about 1 km north of the bridge between Sanibel and Captiva Islands, Florida. The single valves were hand-sorted by size until a collection of several hundred valves, each about 3 cm wide, was available. The greatest grain volume measured in several containers of different shapes and sizes was 29.0%. This value is used as the zero compaction point for curve D (Fig.3-23). Noetia ponderosa is an aragonitic, nearly equivalve species with a deep body cavity, suboval outline, and strong umbos. The loosest packed single-component skeletal sand for which the grain volume was measured is a pebble sand consisting of oval, curved to flat, thin valves of Anomiu simplex Orbigny, the Common Jingle Shell. The shells were handsorted by type and size from the Sanibel-Captiva collection described above. The curved to flat shape of the valve makes it analogous to many types of fossil grain shapes, for example, some of the plate algae. The size of the shells selected for porosity determinations ranged from 1to 3 cm across the valve. Porosity was measured for packs of the Jingle shells in different sized containers and the values were extrapolated to a container of infinite dimensions. Calculated as grain volume, the value entered as the zero compaction point on the Compaction Index chart (curve E , Fig.3-23) is 22.1%. Even looser packing has been reported by Dunham (1962) for red-algal particulate sand (curve F , Fig.3-23) which had a grain volume of 10%. This value is close to that reported by Graton and Fraser (1935) for packed mica and may represent the loosest packing fabric for carbonate particulate sands. Using the Compaction Index chart (Fig.3-23), it can be seen that an oolitic grainstone with a Packing Density of 90% (compressed from 65%)has a Compaction Index of 68%. In other words, this oolitic grainstone has been compacted t o a point where it is 68% of the way from a completely

130

A. H. COOGAN AND R. W. MANUS

uncompacted oolitic sand to a completely compacted rock composed of 100% grains. Oolitic grainstone with a Packing Density of 95% is about 85% compacted (Fig.3-23). On the other hand, with a Packing Density of 65% a grapestone is 35% compacted, a skeletal sand is 42% compacted, a sand of Noetia ponderosa-shaped shells is 51% compacted, a sand of Anomia-shaped particles is 55% compacted and a sand of red algae is 62% compacted (Fig.3-23). More refinements based on shape factors and more determinations of grain volumes are needed to develop widespread applicability of the Compaction Index, but the scheme has the potential of producing comparable values for compaction of limestones regardless of the original differences in packing of the grains. As the Compaction Index is refined and more closely linked to depth of burial, it may be possible to use it to determine the previous depth of burial of now outcropping compacted carbonate grainstones. EXPERIMENTAL STUDIES OF COMPACTION

Summary of experimental work Only a small amount of experimental work has been performed on the compaction of carbonate sands and gravels which could serve as a basis for understanding subsurface compaction of these sediments. In 1966, Fruth et al. reported on the compaction of five Bahamian carbonate sediments subjected to pressures as high as 1,000 kg/cm2 and their work stands as a solid reference point for compaction of carbonate sands. Subsequently, Fruth subjected similar sediments to three types of hydrostatic compaction tests and Ayer (1971) compared the results with ancient compacted grainstones. Earlier, Terzaghi (1940), Hathaway and Robertson (1961), Robertson et al. (1962) and Robertson (1967) reported on the consolidation of calcium carbonate mud. There also has been some interest m the loss of water from high-porosity carbonate sediments on the sea floor under slight consolidation stress from the standpoint of engineering geology (Miller and Richards, 1969). Others have experimented with the compaction of clastic sands (Maxwell, 1960), which have some bearing on understanding the compaction of lime grainstones. Finally, other work on the compaction of spheres (Matzke, 1939; Marvin, 1939) is pertinent to understanding one of the extreme limits of compaction of grain-supported sediment. In this section, the writers first review the experimental work on the compression of spheres, then the pertinent work from papers on the compression of sands in general, and, finally, the experimental compaction of carbonate sands.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

131

Expe rime n ta 1 comp ac tion of spheres Studies on the compaction of spheres are pertinent t o an understanding of the compaction of lime grain sediments, because they provide a limiting condition for the maximum compaction of grains without cement. In an experiment on the compression of randomly-packed, small, spherical lead shot, Marvin (1939) eliminated all void space between particles in a cylinder under pressure of 16,000 kg/cm2. The particles of shot, sized at 2.54 mm (granule size), were compressed at 0.25 cm/min and the shape of the compressed shot and average number of contacts per grain were determined visually under a microscope after discrete stages of compression. The number of grain contacts increased (Table 3-XII) until under a total pressure of 16,000 kg/cmZ, when all space was eliminated, solidly-packed lead polyhedra had an average of 14.16 faces per grain. In a second experiment uniform spherical shot were packed as rhombic regular layers (case 6, Fig.3-5F) and compressed. Each particle, which originally had 6 contact points and 6 free sides facing pores, ended with 1 2 contacts and no new faces, in contrast with the earlier experiment in which new faces were produced in irregular packs. In subsequent experiments, smaller shot (1.27 mm - very coarse sand) were compressed and, except for particles on the periphery of the cylinder, the average contacts per grain for 624 grains of shot were 14.7. Compressed contact faces formed at the original packing contacts were larger than those produced solely from pressure. Neither tamping before compaction nor changing the shot size changed the average of 1 4 contacts per grain after compression. Thus, it can be assumed, that for nearly spherical grains of about uniform size, compaction without concurrent cementation will produce tetrakaidecahedral grains. Mixing proportions of small and large shot (1.27 and 2.54 mm in diameter) in ratios (small to large) of 1:4, 1:l and 4 : l and compressing the mixture under a total load of 18,000 kg/cm2 resulted in different average TABLE 3-XI1 Average number of contacts for 100 observations on spherical, uniform, randomly-packed lead shot compressed under different pressures (After Marvin, 1939)

Pressure (&/ern2 )

Average number of grain contactsper grain

450 2,160 4,530 11,000 16,000

8.41 10.97 12.91 13.62 14.16

132

A. H. COOGAN AND R. W. MANUS

numbers of polyhedral faces. The number of faces per grain for the 1:4 mixture was 9.5 for small shot and 20 for large shot. Equal mixtures had an average of 12 faces for small shot and 19 to 30 for large shot. In the 4 : l mixture, the small shot averaged 10 to 16 faces per grain, whereas the large shot exhibited 25 to 36 faces per grain with the mode of 31 faces. The large shot in the 4:l mix had more contacts, presumably because the large shot was surrounded mostly by small shot. In the equal mix of sizes, some of the large shot, by touching each other, bridge over small shot and protect the latter until compression becomes severe. The number of contacts varied with the size and shape of the particles under compression. The Packing Density for all samples of lead shot at the end of the compression experiments (Marvin, 1939; Matzke, 1939) was 100%. The Packing Index, based on calculations of fitted outlines of irregular dodecahedra, is close to 10076,and the Compaction Index (Coogan, 1970) is 100%.

Experimental compaction of clastic sands - results related to carbonate sands In a series of compaction experiments on quartzose sandstones, Maxwell (1960) attempted to isolate and study independently the following variables: (1) pressure, (2) time, (3) temperature, (4) composition of the sediment, (5) composition of the fluid, and (6) fluid dynamics, i.e., static versus moving fluids. The general impact of the various factors has been considered. In his experiments, the consolidated sands were greatly fractured, interpenetrated and rotated. The larger grains appeared stronger. Most coherent grains showed strain and fractures which are related to points of contact. Solution and redeposition of silica did occur and was enhanced by increased temperature and through-flowing solutions. At pressures equivalent t o a depth of about 9,000 m of overburden, the grain volume was increased to about 70%. Chemical processes were limited to solution, transportation, and precipitation of silica, similar t o the results for carbonate sands. Great emphasis was placed on the importance of fluid movement. For example, Maxwell cited calculations which show that 32,000 cc of water derived from compacting strata have passed through each square centimeter of sediment now found at a depth of 1,750 m in the Cenozoic Ventura Basin of California. Deeply-buried sediments release large amounts of waters saturated with silica (or calcium carbonate, in the case of limestones). This upward moving fluid should lead to upward transportation and deposition of considerable volumes of cement in overlying sediments. It is important to note, however, that cores recovered from the Blake Plateau as part of the JOIDFS program (Schlee and Gerard, 1965; Bunce et al., 1966) contain

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

133

Early Tertiary carbonate sediments, which are buried 150 m (at a depth of 1,180 m below sea level). These sediments are barely consolidated or lithified. Similar uncompacted and very slightly cemented carbonate sediments, with almost normal sea water as connate fluid, were found on the West Florida shelf in the Early and Late Tertiary part of the section. Unlithified and mineralogically unaltered sediments also were found in the Bikini and Eniwetok cores (Schlanger, 1963) recovered from beneath an overburden of over a thousand meters. It seems, therefore, that in these instances there was no significant upward flushing of the fluids. As discussed in the section on experimental compaction of carbonate sediments, fracturing owing t o rapid loading is far more typical of experimentally compacted clastic sands than naturally compacted ones. Pressure solution of grains, development of supersaturated solutions, and precipitation are common where fluids can move. This results in microstylolitic contacts between grains and cemented pore space. In other experiments on clastic sands, Fatt (1958) showed that the compressibility (reciprocal of the bulk modulus) is a linear function of composition for a given grain shape and sorting under pressures up to 1,055 kg/cm2. He chose an idealized model of carbonate composition and texture as a pack of spherical solids with holes in it. He concluded that the effect of the internal fluid pressure on compressibility varies from rock to rock and estimated it to be only 85% effective in counteracting the overburden pressure. According to him, the effectiveness of pore pressure may range from 50 t o 100%. For example, the net overburden pressure at 8,700 m is 1,100 kg/cm2 (based on rock and water density of 2.3 g/cc and 1.0 g/cc, respectively). If the 0.85 effectiveness factor is neglected, then the pressure is 985 kg/cmZ. Based on experimental work and theoretical analysis, however, Rieke and Chilingarian (1974) concluded that pore fluid pressure is 100%effective in counteracting overburden pressure. Compression of sands to 246 kg/cmZ may cause a sharp drop in volume indicating crushing of the grains at that pressure. Plots of grain content versus the bulk volume compressibility (Fatt, 1958) show that intercepts on the y-axis (compressibility) tend to decrease with increasing pressure. This would indicate that for well-sorted silica sands the compressibility decreases with increasing pressure to 844 kg/cm2 and is independent of pressure above that value. This indicates that at very high pressures the' grains in all sandstones are in such close contact and extreme state of strain that the intergranular material does not contribute t o the compressibility. Calcite and quartz have about the same compressibility, but calcite has a very low compressive strength. Compressibility is commonly obtained by a triaxial measurement, or the sample is hydrostatically stressed. Under these conditions, behavior of calcite and quartz is close. When subjected to a

134

A. H. COOGAN AND R. W.MANUS

uniaxial stress, however, the low compressive strength of 2,600 kg/cm2 (compared to 25,300 kg/cm* for quartz) results in crushing of the calcite.

Experimental compaction of carbonate sands Using a triaxial apparatus designed to permit systematic evaluation of pressure, temperature and strain rate, Fruth et al. (1966) compacted carbonate sediments from the Great Bahama Bank which had been selected t o reflect the major sedimentary facies of the Banks. The materials studied by them included sediments from the oolite, oolitic, grapestone, and skeletal facies (Purdy, 1963), all of which are carbonate or muddy carbonate sands. The sediments were subjected to pressures of 1,025 kg/cm2 after determining the initial pore volume. The change in grain volume was calculated from the loss of porosity which was plotted versus increasing confining pressure. In addition, the study included petrographic examination of the compacted sediment for evidence of fracturing, cortical rupture, crushing, and interpenetration of grains. The results of the experiment show that all the sediments became compacted, including the carbonate muds which are not discussed here. The sediments from the oolite facies (Purdy, 1963) consist of polished subspherical ooids and other well-rounded grains, most of which have oolitic coatings (Fig.3-24). Slightly cemented grain aggregates and organic material are common. Size analyses showed a median diameter of 0.35 mm, a Trask sorting coefficient of 1.31, and that less than 0.5% of the sediment was less

Fig.3-24. Oolite facies sand from top of the oolite bar at Browns Cay, Bahamas.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

135

than 0.125 mm in size. The grain volume of the sediment was 60%. This oolite grain sand, with the predominance of solid grains and small amount of mud, was highly fractured by compression. There was negligible initial consolidation because the grains were self-supporting and the fine-grained matrix was very small. Fractures tended to radiate from points of initial contact, but in some cases a single fracture extended completely across the grain (Fig.3-25). The compacted grain volume is 78%. Only a few “floating” grains and rare tangential contacts remained after compression; instead, long contacts predominate. Considerable spalling and buckling of the borders of the ooliths occurred as a result of compaction (Fig.3-25). In spite of the high stress, shells in the oolite sediment are barely affected. Ooliths surrounding a gastropod show spalling and microfaulting of the cortex, whereas a lining of acicular aragonite crystals within the central cavity of the gastropod are hardly affected. Similar spalling may be seen in a compacted oolitic

Fig. 3-25. Artificially compacted Recent carbonate sediments. A. Fractured oolitic grain sand showing marginal peeling and grain penetration after stress of approximately 1,000 kg/cm2. . B. Skeletal facies sediment compacted under approximately 1,000 kg/cm*. (After Fruth et al., 1966, published with permission of the SOC.Econ. Paleontol. and Mineral., J. Sediment. Petrol.)

136

A. H. COOGAN AND R. W.MANUS

grainstone (Fig.3-18 and 3-33),but the intense fracturing of grains occurring under laboratory conditions is seldom duplicated in oolitic rocks compacted in the subsurface. The oolitic facies sediment consists of (1)numerous oolitically coated grains, which are rounded and subspherical; (2) friable and coherent grain aggregates; and (3) skeletal grains with and without oolitic coating (Fig.3-26). The median diameter of grains is 0.35 mm (medium sand), sorting coefficient is 1.4, and about 4% of the sediment is less than 0.125 mm (fine sand) in size. The initial grain volume was approximately equal to 60%. Under stress of 1,025 kg/cm2, the grain volume increased to 84% (CI = 52%). Although fine matrix (4%) tended t o cushion the effects of the stress on grains, the oolitic facies sediment showed abundant grain fracturing. Penetration of grains and buckling of the grain edges are also common. The grapestone sediment (Fig.3-27) consisting of friable and coherent aggregates of skeletal and non-skeletal grains with intraparticulate pore space, had an initial grain volume of about 52%. The grains larger than 0.25 mm in diameter were polished and included well-rounded pellets, shell fragments, and subspherical ooids; the finer grains were more angular. The median diameter is 0.41 mm (medium sand), the sorting coefficient is 1.86, and about 8% of the sediment is below 0.125mm in size. Compaction

Fig.3-26. Oolitic facies sand from the tidal channel between oolite bars at Sandy Cay, Bahamas.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

137

increased the grain volume to 84% grains (CI= 65%), as in the case of the oolitic facies sediment. Among the samples tested, penetration of grains is most pronounced in the grapestone sand. Fracturing at grain contacts was not common, probably because of flowage of soft grains and the abundance of small grains. After stress, all originally “floating” grains were moved into contact. Most of the grains show evidence of penetration, many with minute fractures. The skeletal facies sediment contains abundant skeletal particles of whole foraminifera1 tests, mollusc shells and fragments, and non-skeletal grains of mud and organic aggregates (pellets and pelletoidal grains) (Fig.3-28). The skeletal grains are rounded to angular and are either smooth or rough, whereas the inorganic grains are ovoid. The median diameter is 0.32 mm (medium sand), the sorting coefficient is 2.67, and 22.5% of the sediment is less than 0.125 mm in size. In Dunham’s (1962) classification, this may be called a skeletal packstone sediment as opposed to the grainstone sediments mainly under discussion here. Similar packing of deepsea sediment is shown in Fig.3-29. The original grain volume approximated 3896, which could be accounted for by separation of some of the grains by mud matrix. In addition, the effect of poor shape sorting on the packing also might result in low initial grain volume percentage. After compaction, the grain volume was close t o 88%(CI = 80%)and the principal visual effect was the deformation of the skeletal grains. For example, a moderate-sized shell fractured under stress, whereas the grains contained in the intraparticulate space were not affected.

Fig.3-27. Grapestone facies sand from a grapestone bar top, middle Fish Cay, Berry Islands, Bahamas.

138

A. H. COOGAN AND R. W. MANUS

Fig.3-28. Skeletal facies sand from the burrowed tidal flat near Pigeon Cay, Bimini Lagoon, Bakiamas.

Fig.3-29. Deepsea, planktonic-foraminiferal, packstone-textured sediment from the Tongue of the Ocean, Bahamas.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

139

Fruth et al. (1966) showed that artificial compaction of these carbonate sediments produced many features present in naturally-compacted limestones. Their experiments, however, generally did not produce sediments with textures and fabrics which in composite are similar to those common in compacted lime grainstones. It should be remembered that the experiments were designed to obtain continuous porosity versus pressure curves and the thin sections were made only of the sediment carried to the maximum pressure. It is true that loss of porosity and increase in grain volume accompany both artificial and natural compaction of carbonate sands and that fracturing and grain interpenetration occur in both; however, the mechanism of grain interpenetration during subsurface compaction is through the dual process of pressure solution and mechanical deformation, which results in microstylolitic contacts between the grains. Such contacts were not produced much in these experiments. In addition, grain fracturing is far less common in naturally-compacted grainstones. Carbonate oolitic and grapestone sediments were subjected by Fruth (Ayer, 1971) to three types of hydrostatic compaction* tests. In the first type, the confining pressure was increased in a series of increments and the fluid consisting of reconstituted sea water was allowed to drain freely. This type of test produced structural deformation of the sediment characterized by more sutured contacts, more longitudinal contacts, and less buckling. In the second type of test, the confining pressure was increased at a constant rate and the sea water allowed to drain freely. The result was a compacted sediment with a higher number of concavo-convex contacts, spalling, and diagonal fractures, but an intermediate number of cases of buckling. In the third type of test, the confining pressure and pore pressure were increased simultaneously a t the same rate. After maximum pressures were attained, the pore pressure was decreased at a selected rate. In the three types of experiments, the average maximum confining pressure ranged from about 1,000 to 1,500 kg/cm2. In the third type of test, the compacted sediment had the smallest number of point contacts, fewer concavo-convex contacts than in the type-two test, and the greatest amount of buckling. Comparison of the test materials with naturally-compacted oolitic grainstones showed some differences and similarities. Ayer (1971) concluded, however, that to some extent the predominance of the type of deformation in a naturally-compacted rock was indicative of the depth of burial and might be related to the type of loading. In geological terms, the type-one tests may be similar to the conipaction effects of periodic sedimentation with the concurrent release of pore fluids. The type-two tests might be compared with continuous compaction owing to nearly constant rate of sedimentation and concurrent pore fluid release. The type-three tests *Method described by Fruth et al. (1966).

140

A. H. COOGAN AND R. W. MANUS

might be compared with continuous sedimentation of a sediment confined by impermeable seals which are later broken by faulting. The latter case is probably rare for carbonate sedimentary realms. Petrographic examination shows that the radial and diagonal fractures and the crushing features are restricted to artificial tests and are not found in the compacted oolitic grainstones. The fractures are found in desiccated grainstone sediment, however. In contrast, almost no spalling or buckling features are found in deeply-buried samples and the degree of suturing increases with depth, as reported by Taylor (1950) for clastic sands. In conclusion, it appears that laboratory experiments on the compaction of carbonate sediments have revealed important and useful data, and, while not always directly comparable with naturally-compacted rocks, should be encouraged. Further experimentation involving longer periods of time, more reactive pore fluids, and varied temperature conditions is desirable. COMPACTION AND LITHIFICATION OF CARBONATE SANDS

General problem Experimental works on the compaction of lead shot as well as clastic and carbonate sediments and the consideration of regular packing in sediments provide a necessary theoretical base. They, however, do not elucidate the perplexing facets of compaction of carbonate sands under natural conditions, because of concurrent changes in carbonate mineral species, solution, and cementation of grains. In the simplest case, the lithification of a wholly grain carbonate sediment consisting of nearly spherical particles may be thought of as proceeding in one of three ways (Fig.3-30). Assuming the sediment has an initial grain volume of 65%, the simplest form of lithification would consist of the introduction of cement between the particles of the sediment in an amount equal to that of the initial porosity of 3596, without compaction or solution of the particles. The resultant rock consists of 65%grains, as did the sediment, and the 35% cement is obtained from a non-local source, presumably from solution of another carbonate sediment, or as a precipitate from the pore water (Bathurst, 1971b). A completely lithified beach rock on a modern tropical beach cemented by aragonite (Friedman, 1968) would serve as one example of this type of lithification. Many ancient lime grainstones appear t o have been cemented without compaction or noticeable solution (Fig.3-31-3-33). This widespread and, locally, pervasive cementation doubtless prevents further compaction of the carbonate sands in many otherwise compactive-prone settings.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

141

I 5

m

Fig.3-30. Three lithification processes of spherical grains, shown as a monolayer. A. Original sediment and grains; grain volume = 65%, pore volume = 35%, 6-8 contacts per grain. B. Simple cementation without compaction; grain volume = 65%, cement volume = 35%, 6-8 contacts per grain. C. Complete compaction without cementation; grain volume = loo%, pore plus cement volume = 0%, 8-14 contacts per grain; vertical centerto-center (of spheres) distance reduction = 33%. D. Compaction by pressure solution and cementation; grain volume = 75%, cement volume = 25%, 8 and more contacts per grain; vertical reduction in center-to-center distance = 14%.

A second type of lithification process, starting with the same sediment having a 65% grain volume, would consist of compressing, squeezing and dissolving the grains until they are so compacted that the rock is composed of 100%grains and no cement. Carbonate grainstones lithified in this manner are extremely rare, if they exist at all, except in an intense stage of metamorphism The experimental compaction of lead shot by Marvin (1939) and of carbonate sediments by Fruth et al. (1966) are partial illustrations of this type of compaction. Finally, there is the intermediate case wherein the sediment is compacted &d at the same time cemented, resulting in a rock with some grain volume above the initial uncompacted value of 65% and below the

142

A. H. COOGAN AND R. W.MANUS

Fig. 3-31. Weakly-compacted skeletal oolitic grainstone from the Mississippian Fredonia Limestone Member, Ste. Genevieve Formation, Harrison Co., Ind.

Fig.3-32. Compacted oolitic grainstone from the Jurassic Smackover Formation, showing quartz grain penetrating and breaking carbonate grain. From the North Central Oil Co., J. F. Roper No. 1 well, Freestone Co., Texas, depth = 3,696 m.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

143

Fig.3-33. Compacted oolitic grainstone showing spalling of oolith margins. Collected from the outcropping Union Member, Greenbriar Formation (Mississippian), Monroe Co., W. Va.

theoretical maximum of 100% grains. This appears to be the usual case for compacted lime grainstones.

Sources of cement during compaction of carbonate sands One of the principal unknowns in our understanding of carbonate sand compaction is the source of various cements found in compacted grainstones (Bathurst, 1971b). It is widely observed, for example, that Paleozoic rocks which have not been buried to great depths, such as the Pennsylvanian quartzose sandstones of Kentucky and Indiana, are not much compacted or cemented. Nevertheless, the carbonate rocks interbedded with them or immediately underlying them, for example, the Ste. Genevieve Group oolitic grainstones (Fig.3-31), are cemented. If the cement was precipitated from sea water during the syngenetic stage or from interstitial fluids during the diagenetic stage, the source of such a tremendous volume of cement, representing about 35% of the volume of several cubic miles of rock, can be accounted for with varying degrees of ease. If the cement, on the other hand, must be attributed solely to the solution of grains at the points of contact during compaction, the problem arises as to whether there is sufficient

144

A. H. COOGAN AND R. W. MANUS

material to completely cement the rock without causing large, obvious solution cavities t o form (Matthews, 1967). In a general discussion of the cementation of sediments by carbonate minerals, Mackenzie and Bricker (1971) focused on the composition, chemical behavior and mass transport of some common carbonate cements. From a theoretical standpoint, cement generated through pressure solution and precipitation might have the following characteristics:

(1)Environment of cementation (a) At depth in the sedimentary column. (b) In combination with other previously precipitated cements. ( 2 ) Timing (a) Long after deposition of the sediment. ( 3 ) Cementation process (single or multiple event?) (a) Either as a single process over a long period of time or more commonly as multiple events. ( 4 )Mineralogy (a) Either monomineralic or polymineralic. ( 5 )Source (a) Either locally derived or imported.

The mass of materials dissolved or precipitated in a given carbonate sediment volume has been difficult t o predict when the rock is buried, owing to the unpredictable conditions of temperature, pressure, and solution chemistry. Inasmuch as cements derived directly from stationary pore water can only occupy less than about 1%of the pore volume of a sediment after compaction (Mackenzie and Bricker, 1971) and only 8%can come from the transformation of aragonite to calcite, major cementation requires that cement be transported into the sediment from an outside source. The mechanisms of mass transport of sediment are bulk fluid flow and diffusion, with associated dissolution of sediment particles and pore cementation. The solubility of carbonate minerals in water containing COz decreases with temperature and the solubility trends probably account for one or more phases of mutual replacement of silica and carbonate cements during compaction. The mass transfer of cement involved in burial of a carbonate sediment with its contained water has been calculated by Mackenzie and Bricker (1971) to a depth-of 3.5 km for a region with a geothermal gradient moles CaCO,/1,000 g of of 0.3"C/l,OOO m. It was found to be 3.65 water or about 0.4 g CaC0,/1,000 g HzO, assuming a constant Pco2 of 10-2.5 atmospheres, a pH of 7.4, and equilibrium with calcite. Thus, for every 1,000 cm3 of pore space, about 0.15 cm3 of calcium carbonate will be

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

145

precipitated, resulting in a decrease of pore space of 0.02%. Under these conditions, porosity reduction in a sediment from an initial 30 to 1%would require the flowage of about 1,500 1 of sea water through a 1,000 cm3 volume of sediment. At a flow rate of 700 cm/year, less than 1.5 - lo6 years would be required t o transmit this volume of fluid. Even at much lower flow rates the time required for nearly complete cementation would be small. In regions of higher geothermal gradient, the mass transfer and time required for cementation could be reduced greatly. In an extensive discussion of cementation, Bathurst (1971b) called on three main mechanisms which produce substantial amounts of carbonate cement in rocks: (1)the local dissolution of aragonite, (2) influx of sea water, and (3) pressure solution. He maintains, however, that none of these processes can produce the large amount of cement that is known to be present in limestones. While the authors agree with Bathurst (1971b, p.422) that the exact process may not be clear in specific cases, there is no agreement that the process need be a single one or that it is entirely elusive.

Effect of grain size and shape on cementation The effect of grain size and shape on cementation has been investigated experimentally for quartzose sands (Heald and Renton, 1966) and some of their conclusions apply to carbonate sands. Grain size was found to have an important effect on rates of cementation by freely circulating solutions. Well-sorted coarse sands were cemented faster than fine sands because of their greater permeability. F’iner sands were cemented much more rapidly than coarser ones (by a factor of 2.5), however, where the influx of cementing solutions was the same for both cases. On the basis of area differences, very fine sand theoretically would be cemented 16 times faster than very coarse sand and over 90% of the original porosity in the very coarse sand would be present after all the porosity was eliminated in the very fine sand. In addition, if a bed or lens of coarse sand were surrounded by fine sand, porosity reduction due to influx of cementing solutions would eventually cease in the coarse sand after entry of solutions had been prevented by more advanced cementation of the surrounding fine sand. Rates of cementation were also found to vary with the angularity of the grains. Where influx of cementing solutions is the same, cementation proceeds considerably faster in highly angular sands than in rounded sands of the same grain size, mainly as a result of the greater surface area. Because initial consolidation after sedimentation is greater in angular sands, the combined effect of cementation and compaction would result in much more rapid reduction of porosity in highly angular sands. Since fine sands (for the same particle shapes) tend to be more angular, both the factors of size and

146

A. H. COOGAN AND R. W. MANUS

shape favor the more rapid cementation of fine sands for an equal influx of cement. The more rapid cementation of fine sands would result in more rapid increase in strength of their sand-rock body. Hence, upon burial, compaction, and cementation, fine sand bodies are more likely t o become sufficiently indurated to resist increasing compaction than coarse sand bodies if supplied with the same amount of cement. In many instances one should be able t o observe that either the fine sand is more cemented and the coarser sand is more porous, or that both are equally cemented but that the coarse sand has a greater Packing Density.

Varying cement composition There is a commonly observed variation in mineralogical composition of carbonate cement in compacted grainstones. These differences may be attributed t o relative proportions of aragonite, calcite, high-magnesium calcite, and dolomite which formed during penecontemporaneous diagenesis. Most cement in ancient limestones (Bicker, 1971), however, is lowmagnesium calcite. Several investigators (Bathurst, 1971a; Choquette, 1971; Oldershaw, 1971) have emphasized the differences in form and composition of what appear t o be late and early cements. Commonly, interparticulate pore spaces and intraparticulate skeletal pores are lined with a narrow zone of small prismatic nonferroan calcite crystals which constitute less than 20% of the total cement (Oldershaw, 1971). The remaining space is filled with large, randomly-oriented crystals of ferroan calcite. Petrographic evidence suggests that small nonferroan calcite material was precipitated before compaction, which caused fracturing of shells, whereas the large ferroan calcite was deposited after compaction. The origin of the early cement is postulated t o be penecontemporaneous with sedimentation. The late cement, on the other hand, owes its origin to pressure solution, or mass transfer and migration to the pore site from other dissolving sites. Evidence of very late, partly postinduration cementation was offered by Choquette (1971) for some Ste. Genevieve grainstones in Illinois where the cement is epigenetic ferroan calcite formed in the presence of strongly reducing subsurface waters. Geometrical and packing considerations Evaluation from a geometrical standpoint of the amount of imported versus locally-derived cement produced by pressure solution has been made by Rittenhouse (1971) for clastic sands and part of his study is pertinent to the cementation of carbonate sands. Because the relative amounts of porosity loss due t o solution and cementation vary greatly depending on grain shape, angularity, packing, the direction of applied stress, and grain

147

COMPACTION AND DIAGENESIS OF CARBONATE SANDS TABLE 3-XI11

Loss of porosity by solution versus that by precipitation of cement for four packing models (Data from Rittenhouse, 1971) Packing type cubic Porosity loss from solution (%) 13.4 Porosity loss from cementation (%) 9.2

cubic, rotated 45"

orthorhombic

orthorhombic, rotated 30"

18.4 4.2

10.3 4.2

10.0 4.5

mineralogy, the amount of cement derived from solution of the grains versus the amount of cement imported from other sources must also vary. Calculations for ideal packs of cubic and orthorhombic configurations by Rittenhouse (1971) showed that if the percent of sphere radii in segments of spheres removed by solution was equal to 2896, the solution alone could cause a reduction in porosity of 13.3%(from 47.6 to 34.3%) for cubic packs, if dissolved matter were removed from the system. If the material were entirely precipitated in adjacent pores there would be an additional loss of porosity of 9.1% (from 34 3 to 25.2%). The amounts and ratios of porosity loss owing t o solution and to precipitation vary with the packing (Table 3-XIII). Estimation of the maximum amount of cement that can be derived from solution at points of grain contacts for any given amount of pore space reduction can be made for moderately well-sorted to poorly-sorted, rounded to very angular sands (but not for sands having extremely irregularly shaped carbonate particles) by using the case of regular orthorhombic packing model rotated 30". For example, if the porosity of sand is reduced from 35 to 25%, a reduction of lo%, the maximum amount of cement derived from the solution of grains at their points of contact would be 1.7% according to Rittenhouse (1971) (see Fig.3-34). The application of the same analysis to compacted oolitic grainstones (Fig.3-34-3-39) provides the very interesting results that for deeply-buried samples, there is sufficient cement generated through solution and reprecipitation not only to fill all available pore space but also to supply cement for export (Table 3-XIV; see also Fig.3-31-3-33). These data tend to confirm Coogan's (1970) conclusions that some of the cement in the Smackover (Texas, Louisiana) and Greenbriar (West Virginia) oolitic grainstones is late. The calculations show that the source of the cement can be locally derived by compaction of the oolitic sand to a Packing Density value of 82 to 88%.In the case of the Ste. Genevieve oolitic grainstone (Fig.3-31), however, it is clear that a minimum of 13.6% of the cement present would have to come from sources other than those resulting from

148

A. H. COOGAN AND R. W. MANUS

atzoi

zx

15-

W

z

10-

3

z X

5-

a

z I

I

I

I

I

I

TOTAL PORE RE DUCTIO N, Ya

Fig.3-34. Graph showing the method of determining the maximum amount of cement, which can be derived from solution of grains and subsequent reprecipitation in adjacent pore space for a given reduction of total pore volume. Curve A represents the maxima for any sand and should be used for well-rounded, well-sorted sands. Curve B represents the estimated maxima for sands that are very poorly sorted, extremely well-sorted, or very angular. To use the diagram, enter the amount of pore reduction on the abscissa and read the maximum amount of cement derived from solution and reprecipitation for any particular type of sand on the ordinate (e.g., if original pore space = 35% and the present pore space = 15%, then the total reduction in pore space = 35-15 = 20%;using curve A, the maximum amount of cement = 7.6%). (After Rittenhouse 1971, published with permission of the Am. Assoc. Pet. Geol., Tulsa, Okla.)

Fig.3-35. Uncornpacted Pleistocene eolian oolitic grainstone from the outcrop at Joulters Cay, Bahamas.

149

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

Fig. 3-36. Interpenetrating grains in the compacted oolitic grainstone of the Jurassic Smackover Formation, from Trahan No. 1 Brown well of Tenneco Oil Co., Clairborne Parish, L a , depth = 3,034 m.

TABLE 3-XIV Proportions of solution-derived cement versus total cement volume for selected oolitic grainstones (Fig.3-31-3-33) Sample formation name

Location Fig. no. Original porosity (estimated) Packing density (%) Total cement volume (a) Maximum cement from solution and reprecipitation (%)

Difference between total cement volume and solution-derived cement (excess or deficit) ("5)

Ste. Genevieve

Smackover

Green briar

Illinois (3-31) 35.0

Texas (3-32) 35.0

West Virginia (3-33) 35.0

67.4 32.6 19.0

88.7 11.3 19.0

82.7 17.3 19.0

+7.1

+ 1.7

-13.6

Fig.3-37. Straight contacts, bridging, and grain interpenetration in a compacted oolitic grainstone from J. F. Roper No. 1 well of the North Central Oil Co., Jurassic Smackover Formation, Freestone Co., Texas, depth = 3,696 m.

Fig. 3-38. Three compacted Jurassic, Cretaceous, and Mississippian oolitic grainstones. A, Illustration shows penetration of harder clastic grain into oolith and the authigenic growth of corners into another oolith; from the Jurassic Smackover Formation, North Central Oil Co., J. F. Roper No. 1 well, depth = 3,696 m. B. Illustration shows fractured cement in the Cretaceous El Abra Limestone from the Pemex Cazones No. 2 well, San Luis Potosi, Mexico, depth = 2,421 m. C. Closely-packed, but generally not touching, grains in the Mississippian Greenbriar Formation outcrop in Monroe Co., W. Va.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

151

Fig. 3-39. Two compacted Mississippian and Jurassic ooIitic grainstones. A. Closely-packed grains which are not touching in the Mississippian Ste. Genevieve Formation outcrop, b a n e Co., Tennessee. B. Authigenic crystal growth between compacted ooliths in the Jurassic Smackover Formation, Tenneco Oil Co., No. 1 Waller well, Clairborne Parish, La., depth = 4,022 m.

152

A. H. COOGAN AND R. W. MANUS

local pressure solution. It should be noted, however, that the Packing Density for this rock is close to the value expected for a completely uncompacted oolitic grainstone. This in itself suggests that no compaction has taken place and that all the cement is either of syngenetic or early diagenetic origin, owing to solution and alteration of near-surface sediments. PRESSURE SOLUTION AND STYLOLITIZATION

General statement Pressure solution is important in the compaction of a carbonate sediment because it is an effective agent of bulk volume reduction through the gross dissolution of calcium carbonate, as well as the agent of pore volume reduction by the physical reduction of the void space. Pressure solution also acts as an agent supplying cement to void space adjacent to the site of solution. In addition, it has been suggested by many authors that the calcium carbonate released in the process of pressure solution may serve as a major source of exportable pore-filling cement. Sty lolitiza tion process

Physical features develop in response to dissolution at points of contact between mineral grains subject to pressure. Dissolution occurs as aresult of local increase in chemical potential and thus of solubility of calcium carbonate where grains are subjected to increased elastic strain at the contact points. Where grains are not in contact and the elastic strain is consequently less, the relative solubility of the calcium carbonate is lower and ions diffusing from the more saturated contact areas will tend to precipitate near the unstrained surfaces (Bathurst, 1958, 1971b). The distribution of strain at the surface of contacting grains is a function of their relative sizes, shapes and orientations. Solubility also varies with the mineralogy of the grains and hence the geometry of the contact surface between carbonate grains can be expected to be highly variable. Attempts to classify the geometry of pressure solution features, especially stylolites, have been made by Amstutz and Park (1967) and Trurnit (1968). For the process of pressure solution to continue, stress must be passed across the contact from grain to grain, while a solqtion film is simultaneously maintained between the grains so that dissolved ions may diffuse through it to sites of lower stress. Two mechanisms have been offered to explain the pressure solution: (1)The undercutting mechanism (Bathurst, 1958) calls for the estab-

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

153

lishment of an irregular surface between the grains at a microlevel. Solution is visualized as taking place over only a small part of the surface at one time, undercutting higher relief portions of the surface. (2) The solution film mechanism (Weyl, 1959) calls for the maintenance of a continuous film of solution between the grains, which is able to support shear stress and allow the diffusion of ions. The maintenance of the solution, it is argued, is implied in the force of crystallization noted in experiments (Becker and Day, 1916; Taber, 1916). A stylolite is a complex interface between two bodies of rock, commonly limestone, which exhibits mutual column and socket interdigitation. The long axes of the columns and sockets are perpendicular to the interface and are generally parallel to one another. The interspace between the two bodies of rock separated by a stylolite commonly is characterized by a thin seam of insoluble residue which became concentrated as the rock was dissolved. The difference between microstylolitic grain contacts and those which cut across many grains is simply one of scale. Microstylolitic solution can occur after the precipitation of first generation cement, because a thin fringe of cement does not completely prevent grain-to-grain movement. When the second generation cementation is sufficiently advanced, however, relative movement between grains must end. Consequently, grain-to-grain pressure solution must occur before the complete precipitation of pore-filling cement. Grain-to-grain movement and, hence, pressure solution obviously will be inhibited seriously in cases where appreciable cementation precedes the development of stress conditions sufficient t o produce stylolitization, for example, in cases of vadose zone cementation. Because pressure solution appears to be a response to overburden pressure, it would seem desirable to correlate some function of pressure solution with depth of burial. Unqualified generalizations regarding such relationships, however, appear to be tenuous now. For example, Dunnington (1967), in discussing stylolitic carbonate petroleum reservoirs, indicates that depths of greater than 600 m have been required generally for stylolitization, whereas Schlanger (1964) has described numerous microstylolites from limestones on Guam which were probably never buried more deeply than 90 m.

Bulk volume loss through stylolitization Stylolitization can cause compaction of a sequence of carbonate rocks owing t o substantial losses of rock volume through solution. Reductions in thickness of the order of 30%are common and Dunnington (1967) reported that reductions of up to 40% were recorded. The loss of bulk volume through stylolitization may be diagrammatically illustrated (Fig.3-40);

154 0%

A. H. COOGAN AND R. W. MANUS

,

I---

Flnal V o l u m e

I00 %

60% Original Volume

100 %

Fig. 3-40. Diagrammatic illustration of the bulk volume loss of a limestone through stylolitization.

however, it is very hard to reconstruct lost bulk volume by matching or restoring stylolitic surfaces. If the interpenetration lengths of stylolite columns on each seam are added, a minimum magnitude of unit thinning may be obtained. This must be a minimum value because the degree of solution occurring at the column tips cannot be determined. Another approach to estimating the bulk volume loss is t o cdculate the volume of the insoluble residue from the thickness of the seam and then convert it to bulk volume by using the average insoluble material content of the undissolved rock. The tendency of stylolites to develop in more argillaceous parts of limestones, however, makes this approach inaccurate because of inherent sample bias (Dunnington, 1967). A measure of the volume of rock dissolved in stylolitization may also be based upon the amount of apparent offset of an oblique vein by a cross-cutting stylolite (Pettijohn, 1957). Pore volume loss through pressure solution Pore volume is also reduced by pressure solution as a result of closer packing of grains, grain corrosion, and filling of pores with cement generated by pressure solution. The degree of pore space reduction to be expected by the solution of grains at points of contacts for several ordered packing arrangements of spheres has been demonstrated by Rittenhouse (1971), who also estimated the volume of cement which could be generated by pressure solution. Inasmuch as pore volume equals bulk volume minus grain volume and the original porosities for ordered packing arrangements of spheres are known (Table 3-I), the degree of pore volume reduction can be shown as a function of decrease in bulk volume (Fig.3-41). The data for this model were obtained following the method devised by Rittenhouse (1971) and do not reflect the added influence of void-filling cement, which was locally generated and precipitated through pressure solution. In Fig.3-41, the degree of pore volume reduction (as a percent of original porosity lost) is plotted

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

K 0

155

601

20

40

60

BULK VOLUME L O S S , X

Fig.3-41. Relationship between the percentage of original porosity loss (% 0 Loss) and the percentage of bulk volume loss (% B V Loss) for regular cubic and orthorhombic packs of spheres. A = cubic, B = orthorhombic.

versus the bulk volume reduction owing t o pressure solution for cubic and orthorhombic regular packs of spheres. A reduction of 40% in bulk volume, causes a 35 t o 44% reduction in pore space without cementation, depending on the packing arrangement. Noticeable decrease in porosity has been observed in the immediate vicinity of stylolites (Harms and Choquette, 1965; Dunnington, 1967). Inasmuch as the pore fluids in a region undergoing pressure solution are highly saturated, there is a temptation t o explain this pore reduction as being largely due to cementation. Examination of Fig.3-41, however, reveals that the decrease in porosity is due at least as much t o tighter packing as it is to cementation. An example of tighter packing accompanying a stylolite occurs in an oolitic grainstone from the Silurian Noix Oolite Member of the Edgewood Formation, Louisiana, Missouri (Fig.3-42). Pressure solution as a source of cement

It is frequently suggested that pressure solution may play a significant role in providing a source of allochthonous calcium carbonate which is subsequently precipitated as cement, commonly as second generation cement. Because of the high degree of calcium carbonate saturation of the

156

A. H. COOGAN AND R. W. MANUS

Fig.3-42. Oolitic grainstone from the Silurian Noix Oolite Member of the Edgewood Formation, La., Mo., showing loose and tight packing of ooliths in an area of stylolitization.

pore solutions, i t is not likely that pressure-solution generated cement could travel far before being precipitated in the available pore space. Accordingly, the amount of dissolved calcium carbonate which is being produced must exceed the remaining pore space volume in order for the cement to be exported from the immediate vicinity. Examination of Fig. 3-43, however,

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

157

which shows the relationship between the bulk volume reduction and the ratio of cement generated to the remaining pore volume sheds doubt upon the efficacy of the process. In Fig.3-43 a comparison was made between the data obtained for regular cubic and orthorhombic packs of spheres according to the method used by Rittenhouse (1971). In addition, data was generated (Fig.3-43) for a regular pack of cylinders arrayed so that the ends of the cylinders appear stacked in a cubic regular order. As shown in Fig.3-43, the available pore space exceeds the volume of produced cement until a point is reached where the cement generated ( C G ) divided by the pore volume remaining equals unity. There the bulk volume is reduced by an amount equal t o the original porosity. Indeed, in the case of cubic and orthorhombic packs of spheres (Fig.3-43, curves A, B ) , the bulk volume reduction would have t o be in excess of that commonly supposed to occur before allochthonous cement could be generated. Of course, this assumes local precipitation of cement so long as pore space is locally available. In the case of cubic packs of cylinders (Fig.3-43, curve C), the original porosity (21.5%) is lower and the bulk volume loss required is more in keeping with most estimates of volume reduction. Because it is often considered that ions derived from pressure solution

BULK VOLUME LOSS, K

Fig.3-43. Relationship between bulk volume loss (% BV Loss) and the ratio of cement generated (CG) t o remaining pore volume (6)for regular packs of spheres and cylinders. A = cubic packs of spheres, 47.6% original porosity; B = orthorhombic packs of spheres, 39.5% original porosity; C = cubic packs of cylinders, 21.5% original porosity.

158

A. H. COOGAN AND R. W. MANUS

travel in fluids parallel to the direction of stress (Bathurst, 1971b), it is tempting to suggest that pressure solution is a self-limiting process. There is a reduction in bulk volume until the precipitated cement fills the remaining local pore space so that the permeability is reduced to zero and calcium carbonate can no longer be transported. A t this point pressure solution ceases. The condition may be modified somewhat in the case where cement is transported outside the immediate pore vicinity early in the pressure solution process by rapidly flowing solutions. In this case the curve may rise above the CG/$ ratio of 1.0 (Fig.3-43). In addition, if permeability is increased by the presence of transparticulate pores, such as fractures, this process may be modified. TABLE 3-XV Comparison of Packing Density, Packing Index and Compaction Index of uncompacted and compacted oolitic grainstones

Description ( 1 ) Uncompacted, uncemented oolitic sand, Browns Cay, Bahamas. ( 2 ) Weakly-compacted doloarenite; a dolomitized, weakly-compacted oolitic grainstone showing the relic oolitic texture of grain margins from a depth of 3,917-3,919 m in the Pan American Petroleum Corp., No. 1 Parker, Van Zandt Co., Texas, Jurassic, Smackover Formation. ( 3 ) Compacted oolitic grainstone, Jurassic Smackover Formation from a depth of 3,149 m in the Tenneco Oil Co. No. 1 Lowe, North Haynesville Field, Clairborne Parish, La. ( 4 ) Weaklycompacted, skeletal oolitic grainstone from the M ississipian St e. Genevieve Formation, Fredonia Limestone Member, Harrison County, Ind. (5) Compacted Smackover oolitic grainstone from a depth of 3,696 m in the J. F. Roper No. 1. (6) Compacted oolitic grainstone, Mississippian Greenbriar

Compaction Fig. no. Index (%)

Packing Density (%)

Packing Index (%)

64.8

18.6

0

3-19

70.3

22.5

15

3-20

95.9

66.5

87

3-22

67.4

30.0

7

3-31

88.7

59.5

69

3-32

82.7

31.2

50

3-33

159

COMPACTION AND DIAGENESIS OF CARBONATE SANDS TABLE 3-XV - continued

Description Formation, Union Member, Monroe Co., W. Va. outcrop. Pleistocene eolian oolitic grainstone, uncompacted, Joulters Cay outcrop, Bahamas. Compacted oolitic grainstone, Jurassic Smackover Formation from a depth of 3,034 m in the Trahan No. 1 Brown, Clairborne Parish, La. Compacted oolitic grainstone, Jurassic Smackover Formation from a depth of 3,696 m in the J. F. Roper No. 1. Compacted El Abra Formation; oolitic lump grainstone from a depth of 2,421 m, Pemex Cazones No. 2. Compacted oolitic grainstone, Jurassic Smackover Formation from a depth of 3,696 m in the North Central Oil Co., J. F. Roper No. 1, Freestone Co., Texas. Compacted oolitic grainstone from the Greenbriar Formation, Union Member, Monroe Co., W. Va.; outcrop. Compacted oolitic grainstone, Mississippian Ste. Genevieve Formation, Roane Co., Tennessee; outcrop. Compacted oolitic grainstone, Jurassic Smackover Formation from a depth of 4,022 m in the Tenneco Oil Co., No. 1 Waller, Clairborne Parish, La.

Packing Density (%)

Packing Index (%)

Compaction Fig. no. Index (%)

64.4

21.6

0

3-35

88.3

74.0

66

3-36

84.3

43.1

55

3-37

78.1

34.6

37

3-3813

86.2

46.4

60

3-38a

92.4

23.0

78

3-38~

82.9

18.6

51

3-39a

93.3

74.0

82

3-391,

160

A. H.COOGAN AND R.W.MANUS

GLOSSARY

Chance packing: The natural and fortuitous combination of systematic colonies of particles (chiefly case 6, rhombohedral, Fig.3-5) with intervening or surrounding haphazard or unordered zones of particles. Compaction: The decrease in bulk volume of sediments as a result of stress accompanied by a loss of pore volume. Compaction Index: A measure of the degree of compaction of a grain sediment or rock, scaled in percent, derived from the grain volume measurement and adjusted for the original grain volume percentage of the uncompacted sediment, according t o grain type. Condensation Index: The index of compaction based on the determination of the ratio of fixed to free grains, as determined in thin section. Consolidation: The adjustment of a saturated sediment in response to increased load and consequent loss of water. Fixed Grain: A grain which has its fixed margin (length of the margin of the grain in contact with another grain) longer than its free margin. The measurement is made in thin section. Fixed-Grain Compaction Index: An index of compaction based on the number of fixed grains compared to total grains in the rock. Grain volume: The volume of solid grains in the rock divided by the total bulk volume. Grain volume can be expressed as a fraction or, more commonly, as a percentage of the total rock or sediment bulk volume. Haphazard packing: All packings of spheres which do not correspond to one of the six systematic packing configurations. In haphazard packing, more or less order may prevail, but commonly a disorderly arrangement predominates. Haphazard packing is a special case of random packing. Irreducible wetting-phase saturation: A minimum wetting-phase saturation. Fluid is immobile even at very high pressure and occupies minute interstices and cracks in the sediment or rock and is present as thin films around grains. In most natural systems, the wetting phase is either fresh or salt water. Some carbonate rocks, however, are oil wet. JOIDES: The consortium of oceanographic institutions in the United States of America banded together to investigate the strata of the ocean floors (Joint Oceanographic Institutions for Deep Earth Sampling). Packing: The spacing or density pattern of grains or pores in a sediment or rock. Packing Density: The grain volume percentage of a sediment or rock as measured in thin section. Packing heterogeneity: Variations of packing in a sediment or rock owing to differences in particle size and shape or the presence of alternate regions of variously homogeneously packed particles (see chance packing, haphazard packing, and random packing).

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

161

Packing Index: The index of closeness of grains to each other in a sediment or rock on the average. Packing Index is determined from the ratio of number of grains touching each other t o the total number of grains as counted in thin section. Packing Proximity: One of the f o r m of the Packing Index, specifically that proposed by Kahn (1956). Random packing: A general term to encompass all cases of packing of particles which are not one of the six systematic, regular, ordered packings of spheres. Regular packing: The ordered packing of spheres, one of the six systematic cases of packing (Fig.3-5). Unordered packing: The arrangement of grains in a sediment or rock which does not correspond t o any ordered arrangement or chance packing.

REFERENCES Adams, J. E. and Rhodes, M. L, 1960. Dolomitization by seepage refluxion. Bull. A m . Assoc. Petrol. GepL, 44: 1912-1920. Alderman, A. R. and Skinner, H. C. W., 1957. Dolomite sedimentation in the South-East of South Australia. A m . J. S c i , 255: 561-567. Allen, J. R. L., 1962. Petrology, origin and deposition of the highest lower Old Red Sandstone of Shropshire, England. J. Sediment. Petrol., 32: 657-697. Amstutz, G. C. and Park, W. C., 1967. Stylolites of diagenetic age and their role in the interpretation of the Southern Illinois fluorspar deposits. Mineralium Deposita, 2: 44-53. Athy, L. F., 1930. Density, porosity and compaction of sedimentary rocks. Bull. A m . Assoc. Pet. Geol., 14: 1-24. Atwater, G. I. and Miller, E. E., 1965. The effect of decrease in porosity with depth on future development of oil and gas reservoirs in southern Louisiana. Bull. Am. Assoc. Pet. GeoL, 49: 344. Atwood, D. K and Bubb, J. N., 1970. Distribution of dolomite in a tidal flat environment, Sugarloaf Key, Florida. J. Geol., 78: 499-505. Ayer, N. J., 1971. Statistical and Petrographic Compcrison of Artificially and Naturally Compacted Carbonate Sediments. Thesis, Univ. Ill., Urbana, Ill., 92 pp. Bathurst, R. G. C., 1958. Diagenetic fabrics in some British Dinantian limestones. Liuerp. Munch. GeoL J . , 2: 11-36. Bathurst, R. G. C., 1959. Diagenesis in Mississippian calcilutites and pseudobreccias. J. Sediment. Petrol., 29: 365-376. Bathurst, R. G. C., 1971a. Two generations of cement. In: 0. P. Bricker (Editor), Carbonate Cements. John Hopkins Univ. Press, Baltimore, Md., p.296. Bathuust, R. G. C., 1971b. Carbonate Sediments and Their Diagenesis. Elsevier, Amsterdam, 620 pp. Beall, A. 0.and Fischer, A. G., 1969. Sedimentology. In: Initial Reports o f the Deep Sea Drilling Project, 1. U.S. Gov. Print. Off., Washington, D.C., pp.521-593. Becker, G. F. and Day, A. L., 1916. Note on the linear force of growing crystals. J. Geol., 24: 313-333.

162

A. H. COOGAN AND R. W. MANUS

Bernal, J. D. and Finney, J. L., 1967. Random packing of spheres in non-rigid containers. Nature, 214: 265-266. Benson, L V. and Matthews, R. K., 1971. Electron microprobe studies of magnesium distribution in carbonate cements and recrystallized skeletal grainstones from the Pleistocene of Barbados, West Indies. J. Sediment. Petrol., 41: 1018-1025. Bernal, J. D. and Mason, J., 1960. Coordination of randomly packed spheres. Nature,

188: 910-911.

Bissell, H. J. and Chilingar, G. V., 1967. Classification of sedimentary carbonate rocks. In: G. V. Chilingar, H. J. Bissell and R. W. Fairbridge (Editors), Carbonate rocks. Developments in Sedimentology, 9A. Elsevier, Amsterdam, 47 1 pp. Bishop, W. F., 1968. Petrology of Upper Smackover Limestone in -North Haynesville Field, Clairborne Parish, Louisiana. Bull. A m . Assoc. Pet. Geol., 52: 92-128. Bricker, 0.P. (Editor), 1971. Carbonate Cements. John Hopkins Univ. Press, Baltimore, Md., 376 pp. Brown, R L and Hawksley, P. G., 1945. Packing of regular (spherical) and irregular particles. Nature, 156: 421-422. Bubb, J. N. and Atwood, D. K., 1968. Recent dolomitization of Pleistocene limestone by hypersaline brines, Great Inagua Island, Bahamas. Bull. A m . Assoc. Pet. Geol., 52:522. Bunce, E I., Emery, K. O., Gerard, R. D., Knox, S. T., Lidz, L., Saito, T. and Schlee, J., 1966.Ocean drilling on the continental margin. Science, 150: 709-716. Butler, G. P., 1971. Origin and controls on distribution of arid supratidal (sabkha) dolomite, Abu Dhabi, Trucial Coast. BulL A m Assoc. Pet. Geol., 55: 332-333. Carozzi, A. V., 1961. Distorted oolites and pseudoolites. J. Sediment. Petrol., 31:

262-274.

Chave, K. E,1962. Factors influencing the mineralogy of carbonate sediments. Limnol. Oceanogr., 7: 218-233. Chave, K. E, Deffeyes, K. S., Weyl, P. K., Garrels, R. M. and Thompson, M. E., 1962. Observations on the solubility of skeletal carbonates in aqueous solutions. Science,

137: 33-34.

Chilingar, G. V. and Bissell, H. J., 1961. Dolomitization by seepage refluxion (discussion). Bull. A m Assoc. Pet. GeoL, 45: 679-683. Chilingar, G. V., Mannon, R. W. and Rieke 111, H. H., (Editors), 1972. Oil and Gas Production f r o m Carbonate Rocks. Am. Elsevier, New York, N.Y., 409 pp. Choquette, P. W., 1971. Late ferroan dolomite cement, Mississippian carbonates, Illinois Basin, U.S.A. In: 0. P. Bricker, (Editor), Carbonate Cements. John Hopkins Univ. Press, Baltimore, Md., 376 pp. Coogan, A. H., 1970. Measurements of compaction in oolitic grainstone. J. Sediment. Petrol., 40: 921-929. Coogan, A. H., Bebout, D. G. and Maggio, C., 1972. Depositional environments and geologic history of the Golden Lane and Poza Rica Trend, Mexico, an alternative view. Bull. A m Assoc. Pet. GeoL, 56: 1419-1447. Deffeyes, K. S., Lucia, F. J. and Weyl, P. K., 1965. Dolomitization of Recent and Hio-Pleistocene sediments by marine evaporite waters on Bonaire, Netherlands Antilles. In: L. C. Pray and R. C. Murray (Editors), Dolomitkotion and Limestone Dingenesis - A Symposium Soc. Econ Paleontol. Mineral., Spec. Pap., 13: 71-88. Dunham, R. J., 1962. Classification of carbonate rocks according to depositional texture. In: W. E Ham (Editor), Classification o f Carbonate Rocks - A Symposium. A m. Assoc. Pet. GeoL, Spec. Publ., 14: 108-121. Dunnington, H. V., 1967. Aspects of diagenesis and shape change in stylolitic limestone reservoirs. Proc. World Pet. Congr., 7th., Mexico, 1967, 2: 339-352.

COMPACTION AND DIAGENESIS OF CARBONATE s ANDS

163

Ellis, A. L,. and Lee, C. H., 1919. Ground waters of western San Diego County, California. U.S.G.S. Water Supply Pap., 446: 121-123. Fatt, I, 1958. Compressibility of sandstones at low to moderate pressures. Bull. A m . Assoc. P e t Geol., 42: 1924-1957. Fischer, A. G. and Garrison, R. E., 1967. Carbonate lithification on the sea floor. J. GeoL, 75: 488-496. Folk, R. I,., 1959. Practical petrographic classification of limestones. Bull. A m . Assoc. P e t GeoL, 43: 1-38. Fraser, H. J., 1935. Experimental study of porosity and permeability of clastic sediments. J. GeoL, 43: 910-1010. Freund, J. E. and Williams, F. J., 1958. Modern Business Statistics. Prentice Hall, Englewood, N.J., 539 pp. Friedman, G. M., 1964. Early diagenesis and lithification in carbonate sediments. J. Sediment. Petrol., 34: 777-813. Friedman, G. M., 1965. Occurrence and stability relationships of aragonite, highmagnesian calcite, and low-magnesian calcite under deep-sea conditions, Bull. Geol. SOC.A m . , 76: 1191-1196. Friedman, G. M., 1968. The fabric of carbonate cement and matrix and its dependence on the salinity of’water. In: G. Miiller and G. M. Friedman (Editors), Recent Developments in Carbonate Sedimentology in Central Europe. Springer, Berlin, pp.11-20. Ruth, L. S., Jr., Orme, G. R. and Donath, F. A., 1966. Experimental compaction effects in carbonate sediments. J. Sediment. Petrol., 36:- 747-754. Fuchtbauer, H., 1961. Zur Quarzneubildung in Erdoellagerstaetten. Erdol. Kohle, 14 : 169-173. Gaither, A., 1953. A study of porosity and grain relationships in experimental sands. J. Sediment. PetroL, 23: 180-195. Graton, L. C. and Fraser, H. J., 1935. Systematic packing of spheres with particular relation t o porosity and permeability. J. GeoL, 35: 785-909. Ham, W. E. (Editor), 1962. Classification o f Carbonate Rocks - A Symposium. Am. Assoc. Pet. Geol., Tulsa, Okla., 279 pp. Harms, J. C. and Choquette, P. W-., 1965. Geologic evaluation of gamma-ray porosity device. Trans. Soe. Prof Well Log Analysts, Annu. Logging Symp. 6 t h Dallas, Texas, 1965, C, pp.1-37. Harris, C. C. and Morrow, N. R., 1964. Pendular moisture in packing of equal spheres. Nature, 207: 706-708. Hathaway, J. C. and Robertson, E. C., 1961. Microtexture of artificially consolidated aragonitic mud. U.S. GeoL Surv., Prof Pap., 424C: 301-304. Hays, F. R., 1951. Petrographic Analysis o f Deep Well Cores. Thesis, Dep. Geol., Univ. Cincinnati, Cincinnati, Ohio (not seen). Heald, M. T. and Renton, J. J., 1966. Experimental study of sandstone cementation. J. Sediment Petrol., 36: 977-991. Huang, T. and Pierce, J. W., 1971. The carbonate minerals of deep-sea bioclastic turbidites, southern Blake Basin. J. Sediment Petrol., 41: 251-260. Kahle, C. F., 1966. Some observations on compaction and consolidation in ancient oolites Compass, 44: 19-29. Kahn, J. S., 1956. Analysis and distribution of the properties of packing in sand-size sediments, I. On the measurement of packing in sandstones. J. Geol., 64: 385-395. Kelvin, Lord, 1887. On the division of space with minimum partitional area. Phil. Mag., 24: 503-514.

164

A. H. COOGAN AND R. W. MANUS

King, F. H., 1898. Principles and conditions of the movements of ground water. US. Geol. Surv., 19th Annu. Rep., 3: 208-218. Land, L. S. and Goreau, T. F., 1970. Submarine lithification of Jamaican reefs. J. Sediment. Petrol., 40: 457-462. Larsen, G. and Chilingar, G. V., 1967. Diagenesis in Sediments. Developments in Sedimentology, 8. Elsevier, Amsterdam, 551 pp. LeBlanc, R. J. and Breeding, J. G. (Editors), 1957. Regional Aspects of Carbonate Deposition. Soc. Econ. Paleontol. Mineral., Spec. Pap., 5: 1-178. Mackenzie, F. T. and Bricker, 0. P., 1971. Cementation of sediments by carbonate minerals. In: 0. P. Bricker (Editor), Carbonate Cements. John Hopkins Univ. Press, Baltimore, Md., 376 pp. Maiklem, W. R., 1968. Some hydraulic properties of bioclastic carbonate grains. Sedimentology, 10: 101-109. Marvin, J. W., 1939. The shape of compressed lead shot and its relation to cell shape. A m . J. Bot., 26: 280-288. Masson, P. H., 1951. Measurement of grain packing in sandstone. A m . Assoc. Pet. Geol., Annu. Meet., St. Louis, not seen (abstract). Matthews, R. K., 1966. Genesis of Recent lime mud in southern British Honduras. J. Sediment. Petrol., 36: 428-454. Matthews, R. K , 1967. Diagenetic fabrics in biosparites from the Pleistocene of Barbados, West Indies. J. Sediment. Petrol., 37: 1147-1153. Matthews, R. R,1968. Carbonate diagenesis: equilibration of sedimentary mineralogy to the subaerial environment; coral cap of Barbados, West Indies. J. Sediment. Petrol., 38: 1110-1119. Matzke, E. B., 1939. Volume-shape relationships in lead shot and their bearing on cell shape. A m . J. Bot., 26: 288-295. Maxwell, d. C., 1960. Experiments on compaction and cementation of sand. Geol. SOC. Am., M e m , 79: 105-132. Maxwell, J. C., 1964. Influence of depth, temperature and geologic age on porosity of quartzose sandstone. Bull. Am. Assoc. Pet. Geol., 48: 697-709. Milliman, J. D., 1966. Submarine lithification of carbonate sediments. Science, 153: 994-9 9 2. Miller, D. G., Jr. and Richards, A. F., 1969. Consolidation and sedimentation - compression studies of a calcareous core, Exuma Sound, Bahamas. Sedimentology, 12: 301-316. Morrow, N. R., 1971. Small-scale packing heterogeneities in porous sedimentary rocks. Bull. A m . Assoc. Pet. Geol., 55: 514-522. Murray, R. C., 1969. Hydrology of south Bonaire, Netherlands Antilles - A rock selective dolomitization model. J. Sediment. Petrol., 39: 1007-1013. Oldershaw, A. E., 1971. The significance of ferroan and non-ferroan calcite cements in the Halkin and Wenlock Limestones (Great Britain). In: 0. P. Bricker (Editor), Carbonate Cements. John Hopkins Univ. Press, Baltimore, Md., pp.225-229. Park, Won-choon and Schot, E. H., 1968. Stylolitization in carbonate rocks. In: G. Miiller and G. M. Friedman (Editors), Recent Developments in Carbonate Sedimentology in Central Europe. Springer, Berlin, pp.66-74. Pettijohn, F. J., 1957. Sedimentary Rocks. Harper and Row, New York, N.Y., 2nd ed., 718 pp. Pray, L. C. and Murray, R. C. (Editors), 1965. Dolomitizntion and Limestone Diagenesis - A Symposium. Soc. Econ. Paleontol. Mineral., Spec. Publ., 13: 1-180.

COMPACTION AND DIAGENESIS OF CARBONATE SANDS

165

Pryor, W. A., 1971. Reservoir inhomogeneities of some Recent sand bodies. Soc. Pet. Eng., A m . Inst. Min., Metall. Petrol. Eng., Preprint Paper SPE 3607, presented at Annu. Meet., 1971, 1 2 pp. Purdy, E. G., 1963. Recent calcium carbonate facies of the Great Bahama Bank. 2. Sedimentary facies. J. GeoL, 71: 472-497. Purdy, E. G., 1968. Carbonate diagenesis: an environmental survey. Geol. Romana, 7: 183- 2 2 8. Ridgway, K. and Tarbuck, K. J., 1967. Random packing of spheres. Br. Chem. Eng., 12: 384-3 88. Rieke 111, H. H. and Chilingarian, G. V., 1974. Compaction of Argillaceous Sediments. Elsevier, Amsterdam, 424 pp. Rittenhouse, G., 1971. Pore-space reduction by solution and cementation. Bull. A m . Assoc. Pet. GeoL, 55: 80-91. Robertson, E. C., 1967. Laboratory consolidation of carbonate sediment. In: A. F. Richards (Editor), Marine Geotechnique. Univ. 111. Press, Urbana, Ill., pp.118-127. Robertson, E. C., Sykes, L. R. and Newell, M., 1962. Experimental consolidation of calcium carbonate sediment. U.S. Geol. Surv., Prof Pap., 350: 82-131. Robinson, R. B., 1967. Diagenesis and porosity development in Recent and Pleistocene oolites from southern Florida and the Bahamas. J. Sediment. Petrol., 37: 355-364. Scott, G. D., 1960. Packing of equal spheres. Nature, 188: 908-909. Schlanger, S. O., 1963. Subsurface geology of Eniwetok Atoll. U.S. Geol. Surv., Prof. Pap., 260 BB: 991-1066. Schlanger, S. 0..1964. Petrology of the limestones of Guam. U.S.Geol. Surv., Prof. Pap., 403 D: 1-52. Schlee, J. and Gerard, R., 1965. Cruise report and preliminary core log M/V Caldrill I - 17 April to 17 May, 1965. J.O.I. D.E.S. Blake Panel R e p . , 64 pp., unpublished. Scholle, P. A , 1971. Diagenesis of deep-water carbonate turbidites, Upper Cretaceous Monte Antola flysch, northern Apennines, Italy. J. Sediment. Petrol., 4 1 : 23 3- 2 50. Sharp, W. E. and Kennedy, G. C., 1965. The system CaD-C0,-H, 0 in the two-phase region calcite + aqueous solution. J. Geol., 73: 391-403. Shinn, E. A., 1964. Recent dolomite, Sugarloaf Key. In: Guidebook for Field Trip No. 1, GeoL Soc. A m , Conv. 1964, pp.62-67. Shinn, E. A., Ginsburg, R. N. and Lloyd, R. M., 1965. Recent supratidal dolomite from Andros Island, Bahamas. In: L. C. Pray and R. C. Murray (Editors), Dolomitization and Limestone Diagenesis - A Symposium Soc. Econ Paleontol. Mineral., Spec. PubL, 13: 112-123. Sippel, R. F. and Glover, E. D., 1964. The solution alteration of carbonate rocks, the effects of temperature and pressure. Geochim Cosmochim. Acta, 28: 1401-1417. Stearns, N. D., 1927. Laboratory tests on physical properties of water-bearing materials. U S . GeoL S u m , Water Supply Pap., 596 F : 163-169. Taber, S., 1916. The growth of crystals under external pressure. A m . J. Sci., 4: 532-556. Taft, W. H., 1968. Yellow Bank, Bahamas. A study of modern marine carbonate lithification. Bull A m Assoc. Pet. Geol., 52: 551. Taft, W. H. and Harbaugh, J. W., 1964. Modern carbonate sediments of South Florida, Bahamas and Espiritu Santo Island, Baja California: A comparison of their mineralogy and chemistry. Stanford Univ. P u b l , Univ. Ser., Geol. Sci., 8(2): 1-133. Taylor, J. M., 1950. Pore-space reduction in sandstone. Bull. Am. Assoc. Pet. Geol., 34: 7 01-71 6.

166

A. H. COOGAN AND R. W. MANUS

Terzaghi, K., 1936. Simple tests determine hydrostatic uplift. Eng. News-Record, 116:

872-875.

Terzaghi, R., 1940. Compaction of lime mud as a cause of secondary structure. J. Sediment. P etr o l, 10: 78-90. Thompson, G., Bowen, V. T., Melson, W. G. and Cifelli, R., 1968. Lithified carbonates from the deep sea of the equatorial Atlantic. J. Sediment. Petrol., 38: 1305-1312. Trurnit, P., 1968. Pressure-solution phenomena in detrital rocks. Sediment. Geol., 2:

89-114.

Weller, J. M., 1959. Compaction of sediments. Bull A m . Assoc. Pet. Geol., 43: 273-310. Weyl, P. K., 1959. Pressure solution and the force of crystallization - a phenomenological theory. J. Geophys. Res., 64: 2001-2025. Zaripov, 0. G. and Prozorovich, G. El, 1967. 0 razlichiyakh epigenetischeskogo uplotneniya vodonasyshchennykh i neftenasyshchennykh porod-kollektorov (na primere produktivnykh gorizontov Surgutskogo svoda, Zapadnaya Sibir’). (Concerning the differences between the epigenetic consolidation of watersaturated and oil-saturated reservoir rocks, using as an example the productive horizons of the Surgut anticline, Western Siberia.) Dokl. Akad. Nauk S.S.S.R.,

176: 1131-1133.